VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Dr. Cecilia Dones - Founder and Chief Data Officer, 3 Standard Deviations

Dr. Cecilia Dones led an interactive workshop examining how New Yorkers interpret data, how civic data is created through lived experience, and how generative AI systems can misrepresent or overextend conclusions drawn from public datasets. Using NYC 311 complaint data as a case study, she explored how critical thinking skills can help residents evaluate claims generated by dashboards, reports, and AI systems.

Data Literacy Through Everyday Experience

Dones opened the session by framing the discussion around trust, authenticity, and technology-mediated communication. She explained that her research background focuses on understanding whether information encountered through digital systems can actually be trusted. Rather than presenting a technical workshop, she emphasized that the session would focus on practical reasoning and lived experience rather than coding or advanced statistics.

Participants discussed what they had recently noticed in their neighborhoods, mentioning changing weather, flowers, scaffolding, traffic, noise, litter, and trash. Dones then shifted the conversation toward annoyance and reporting behavior, asking attendees whether they actually reported these issues to 311. Most participants acknowledged that although they notice many problems, they often do not formally report them.

From this, Dones introduced one of the workshop’s foundational ideas: not everything becomes data. Data, she argued, is ultimately a human choice about what gets measured, recorded, categorized, and stored. Many aspects of urban life never enter official datasets at all because people choose not to report them, do not know how to report them, or decide the issue is not serious enough.

Thresholds, Reporting, and Selection Bias

The discussion explored how different people have different “thresholds” for deciding when an issue deserves formal reporting. Participants debated whether loud music, trash, or lack of heat should always trigger complaints, eventually concluding that context matters heavily. Noise at noon differs from noise at 3 a.m.; no heat matters differently in winter than in summer.

Dones used these examples to explain that civic datasets are shaped by human tolerance, awareness, and behavior. What gets reported reflects not only the existence of a problem but also the point at which people feel compelled to act. She connected this idea to selection bias, emphasizing that datasets reflect social behavior as much as objective conditions.

She also highlighted awareness as a factor influencing reporting patterns. Some people may not know that 311 exists or may not trust that reporting will produce results. Others may solve problems informally or simply tolerate recurring issues. Missing data therefore also carries meaning because absence from a dataset does not necessarily indicate absence of a problem.

Understanding NYC 311 Data

Dones walked participants through the NYC 311 portal, describing it as a non-emergency system for reporting neighborhood issues such as noise, trash, heat, blocked driveways, and illegal parking. She explained that once complaints are submitted, they are categorized and routed to appropriate city agencies for action.

She stressed the importance of classification systems in civic data. Categories such as “noise” may contain very different phenomena, including construction, traffic, parties, barking dogs, or other disturbances. The way governments define and group these categories shapes how policymakers interpret urban problems and allocate resources. Oversimplified classification schemes can obscure important distinctions and potentially lead to ineffective policy responses.

Drawing on NYC Open Data, Dones reviewed some of the city’s most common 311 complaints, including residential noise, heat and hot water problems, illegal parking, blocked driveways, and other everyday quality-of-life concerns. She emphasized that these datasets are dynamic and continuously changing, noting that 311 information is refreshed daily and reflects evolving social conditions.

Temporal Context and Changing Data

A major theme of the workshop was that data is not static. Dones argued that many people mistakenly assume that official numbers in reports or dashboards represent timeless truth, when in reality they are snapshots shaped by changing circumstances.

To illustrate this, she asked participants how a five-day heat wave might affect complaint patterns. Attendees suggested increases in reports related to cooling centers, power outages, garbage smells, rats, pools, and health issues. Others noted that people might generally become less tolerant and report more issues simply because they were uncomfortable.

Dones connected this to temporal bias — the mistaken assumption that a dataset reflects permanent conditions rather than specific moments in time. She cited a 2019 study of 311 complaints about homelessness, where a sharp increase in reports initially appeared alarming. Researchers later determined that the increase reflected greater public awareness due to a new city program, rather than an actual rise in homelessness itself.

This example demonstrated how shifts in awareness, outreach, or reporting mechanisms can alter datasets without necessarily reflecting underlying social change. Context, therefore, becomes essential for interpreting numbers responsibly.

Generative AI and Misleading Confidence

The second half of the workshop focused on how generative AI systems interact with civic data. Dones narrowed the discussion specifically to large language models (LLMs) such as Claude and ChatGPT, explaining that these systems generate probabilistic responses based on patterns in data rather than direct understanding or reasoning.

She demonstrated this by querying Claude about the most common 311 complaints in New York City. Although the system correctly identified noise as a major issue, it relied on 2024 information rather than current data, despite the fact that 311 datasets update daily. This revealed one limitation of LLMs: their outputs may appear current and authoritative while actually relying on stale information.

Dones then drilled down into increasingly specific geographic questions about Bushwick and her former ZIP code. Claude initially admitted limitations and recommended consulting direct sources such as the NYC 311 monitoring tool and NYC Open Data portal. She viewed this transparency positively because the model acknowledged uncertainty and directed users toward primary sources.

ChatGPT, by contrast, attempted to provide more direct answers even when the underlying data was weak or unavailable. Although its responses sounded plausible and aligned with general assumptions about neighborhood noise, the actual ZIP-code-level data from the 311 monitoring tool showed that heat and hot water complaints were more prominent than noise complaints during the examined period.

This mismatch illustrated a core danger of generative AI systems: they can produce coherent and persuasive narratives that overgeneralize beyond what the underlying data actually supports. Dones emphasized that users must critically evaluate which parts of an AI-generated answer are grounded in real evidence and which parts reflect extrapolation or probabilistic guesswork.

Missing Data and Invisible Stories

The workshop also examined problems that never appear in official data at all. Participants mentioned issues such as blocked bike lanes, film crews, triple parking, unloading trucks, and temporary disruptions that residents often tolerate without reporting.

Dones argued that missing data is itself meaningful. The absence of reports may indicate resignation, lack of awareness, distrust in institutions, or acceptance of recurring problems rather than the nonexistence of those problems.

She revisited earlier research on 311 reporting patterns, noting that increased reporting by city inspectors had significantly altered datasets because inspectors became a new source of complaints within the system. This demonstrated that datasets are shaped not only by social conditions but also by institutional processes and changes in reporting infrastructure.

Building Critical Data Skills

Dones concluded by encouraging participants to develop habits of skepticism and contextual inquiry whenever encountering data, dashboards, reports, or AI-generated analysis. Before trusting a narrative, she recommended asking several questions:

• Where did the data come from?

• Why did it show up in this way?

• What is missing from the dataset?

• What was happening at the time?

• Does the conclusion go beyond what the data actually supports?

She summarized five major lessons from the workshop:

• Not everything becomes data.

• Data reflects when people reach their limit.

• Reporting patterns depend heavily on context.

• AI-generated answers can begin from real data but still overextend conclusions.

• Missing information is itself meaningful and deserves interpretation.

Dones closed by stressing that critical thinking about data will become increasingly important as AI systems and algorithmic tools become more integrated into everyday life. She encouraged participants to continue exploring NYC Open Data, experiment with civic datasets, and develop stronger habits of inquiry rather than accepting official-looking numbers or AI outputs at face value.

RESOURCES

• Can You Trust the Story? — Dr. Cecilia Dones’s NYC Open Data Week workshop page

• NYC311 — the city’s portal for reporting non-emergency issues

• 311 Service Requests dataset — the daily-updated source on NYC Open Data used in the talk

• NYC311 Monitoring Tool — NY State Comptroller dashboard for neighborhood and ZIP-code complaint breakdowns

• NotebookLM — Google’s source-grounded summarization tool Dones cited for working with text

• Claude — one of the two LLMs Dones queried about 311 complaints in the live demo

• ChatGPT — the second LLM compared in the demo on neighborhood complaint data

• 3 Standard Deviations — Dr. Cecilia Dones’s data and AI consultancy

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Alex Richey - Data Engineer, NYC Department of City Planning; Damon McCullough - Data Engineering Team Lead, NYC Department of City Planning

The session explored how the NYC Department of City Planning (DCP) is integrating recent “City of Yes for Housing Opportunity” legislation into PLUTO, the city’s flagship land use and tax lot dataset. Alex Richey and Damon McCullough described both the policy context behind the changes and the technical and organizational challenges involved in translating zoning law into usable geospatial data.

PLUTO, Open Data, and the City of Yes

Alex Richey explained that DCP serves both planning and research functions for New York City and maintains many widely used geographic datasets, including PLUTO, facilities data, and housing databases. The data engineering team relies heavily on Python, PostGIS, dbt, and open source tooling, and publishes many of its workflows publicly.

PLUTO, which stands for “Primary Land Use Tax Lot Output,” contains geographic and land use information for nearly one million tax lots citywide and is updated monthly with larger quarterly releases. The new work focuses on adding fields related to “City of Yes” housing reforms, specifically transit zones and Mandatory Inclusionary Housing (MIH) areas.

The speakers framed the work within New York City’s housing crisis. The recently adopted “City of Yes for Housing Opportunity” legislation seeks to encourage housing development, including by reducing parking minimums and expanding affordable housing requirements. The new PLUTO fields are intended to help planners, researchers, developers, and analysts evaluate housing opportunities and zoning requirements at scale rather than lot-by-lot through existing tools like Zola.

Transit Zones and Parking Requirements

Damon McCullough introduced the zoning concepts underlying the new fields. He explained that transit zones determine parking requirements for new development and are based on four broad geographic categories across the city: the Manhattan Core, Long Island City zone, inner transit zone, outer transit zone, and areas outside those designations. In some zones, parking is not required for residential development, while in others parking requirements still apply.

Alex Richey then described how the team attempted to assign every tax lot in PLUTO to the correct transit zone programmatically. Initially, this appeared straightforward: overlay transit zone polygons with tax lot geometries and determine where each lot belongs. However, numerous edge cases emerged once the process was scaled to nearly a million lots.

The team discovered many lots that straddled transit zone boundaries. Some were simple cases where most of a lot clearly fell within one zone, but others were split almost evenly between two zones. To address this, the team incorporated the concept of tax blocks, reasoning that neighboring lots within the same block could provide contextual clues for assigning ambiguous lots.

This introduced new complications because Department of Finance blocks often did not correspond to intuitive city blocks. Some blocks stretched across waterways or consisted of disconnected geographic pieces. The team therefore created a process to split official tax blocks into contiguous “real” blocks before evaluating their transit zone assignments.

Even this refinement produced difficult cases, including rail corridors and oddly shaped transit boundaries. Ultimately, the engineers developed a decision tree: if a contiguous block fit clearly within a transit zone, all associated lots inherited that assignment; if not, assignments reverted to individual lot-level calculations. This process eliminated the need for manual correction lists and left only a single unresolved outlier lot in Queens that was perfectly bisected by transit zone boundaries.

Mandatory Inclusionary Housing (MIH) Areas

McCullough then shifted to Mandatory Inclusionary Housing, a program adopted in 2016 requiring affordable housing in areas designated for residential growth. He explained the four MIH options, each of which specifies different affordable housing percentages and income thresholds based on Area Median Income (AMI).

The challenge for PLUTO was determining which MIH areas and options applied to each tax lot. While some lots fell cleanly within a single MIH polygon, others intersected multiple MIH areas or had buildings partially outside the designated boundaries.

The team applied PLUTO’s existing rule for spatial overlays: if more than 10 percent of a lot intersects a zone, the lot is considered part of that zone. This created additional ambiguities for lots with exactly 10 percent overlap or multiple overlapping MIH options.

The MIH source data itself introduced inconsistencies, including different spellings and formatting for option names. The engineers initially considered adding only one MIH field to PLUTO, but discussions with housing experts and data users led them to implement four separate boolean-style flag columns indicating whether a lot is associated with MIH Options 1, 2, 3, or 4.

The speakers emphasized that these choices reflected tradeoffs between usability and dataset complexity. They concluded that users cared more about quickly identifying applicable MIH options than about preserving every source attribute such as project names or ULURP numbers.

Open Source Tools and Geospatial Workflows

Throughout the presentation, both speakers highlighted the importance of open source tooling and collaborative workflows. They described using Python, PostgreSQL/PostGIS, QGIS, DBeaver, and internal visualization tools to investigate problematic lots and zoning boundaries.

The presenters stressed that visual inspection of geospatial data was essential for debugging. Looking at maps often revealed problems that would be difficult to detect through SQL queries or tabular inspection alone. They also discussed the practical realities of maintaining large-scale citywide datasets, noting that some ambiguity is unavoidable when combining imperfect source data, zoning law, and automated transformations.

They repeatedly emphasized that PLUTO is not a legal determination of zoning status, but rather a highly useful planning and research dataset. For definitive answers, users still need to consult the Department of Finance or the relevant zoning authorities.

Collaboration, QA, and Lessons Learned

The speakers described the project as highly collaborative, involving frequent discussions with transportation planners, housing specialists, and zoning experts. Rather than attempting to interpret zoning text independently, the engineers regularly brought edge cases to subject matter experts for clarification.

They distinguished between “development QA,” which involves intensive collaboration while building new features, and “routine QA,” which supports the ongoing monthly release process. The team tries to avoid manual correction workflows because those processes do not scale well across hundreds of thousands of lots and recurring monthly releases.

Several broader lessons emerged from the project:

• Large-scale civic datasets inevitably involve tradeoffs between precision and scalability.

• Source geospatial data often contains ambiguities or inconsistencies that become apparent only during implementation.

• Open communication with domain experts is essential for turning legislation into operational data systems.

• Open data becomes more valuable when users publicly share analyses, workflows, and feedback that can improve future releases.

During the Q&A, the presenters encouraged developers and researchers to experiment with NYC Open Data, contribute ideas through open source channels, and consider working directly for city agencies if they want to improve public data infrastructure. They also noted the growing popularity of QGIS among newer planners and GIS practitioners because of its accessibility and open source ecosystem.

RESOURCES

• PLUTO & MapPLUTO — DCP’s tax-lot dataset that the team adds the new fields to

• PLUTO on NYC Open Data — the public dataset (nearly a million tax-lot records), released monthly

• NYCPlanning/data-engineering — the open-source repo where the team builds PLUTO in Python, PostGIS, and dbt

• ZoLa — NYC Zoning & Land Use Map — the sister tool used throughout the talk to inspect individual lots

• City of Yes for Housing Opportunity — the citywide zoning amendment driving the new PLUTO fields

• Transit Zones — the Open Data source layer behind the new transit-zone field

• Mandatory Inclusionary Housing — the Open Data MIH-areas layer behind the new MIH option flags

• Mandatory Inclusionary Housing (HPD) — background on the MIH program adopted in 2016

• Open Data Week event page — the session listing with speaker and dataset links

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Arianna Zuanazzi - Open Science and Research Collaboration Specialist, Child Mind Institute; Adam Santorelli - Software Engineer, Child Mind Institute; Nathalia Bianchini Esper - Software Engineer, Child Mind Institute
Moderator: Arianna Zuanazzi - Child Mind Institute

Introduction to the Child Mind Institute and the Healthy Brain Network

Arianna Zuanazzi introduced the Child Mind Institute as an independent nonprofit organization focused on transforming the lives of children and adolescents facing mental health and learning challenges through care, education, and scientific research. She explained that the presenting team works within the Institute’s science department, which focuses heavily on neuroscience research, open science initiatives, data sharing, and applied technology development.

Zuanazzi emphasized the scale of mental health challenges among children and adolescents:

• Globally, approximately one in five children experience mental health or learning challenges

• Around 70% of U.S. counties lack a child and adolescent psychiatrist

• The average delay between symptom onset and treatment exceeds eight years

To address these gaps, the Child Mind Institute launched the Healthy Brain Network (HBN) initiative in 2015. HBN uses a community-referred recruitment model in which families participate because they have concerns about their child’s mental health. Participants receive free evaluations and recommendations, while researchers gain access to a large open scientific dataset.

The Healthy Brain Network has now collected data from more than 8,000 children and adolescents, primarily in New York City, creating one of the richest openly shared pediatric mental health datasets available.

Characteristics of the Healthy Brain Network Dataset

Zuanazzi reviewed the demographic and diagnostic composition of the HBN sample.

Participants range from:

• 5 to 22 years old

• With the majority between ages 6 and 11

Because the sample is community-referred rather than randomly selected, more than 90% of participants have received at least one diagnosis.

The most common diagnoses include:

• ADHD (over 70%)

• Anxiety disorders (approximately 50%)

• Learning disorders (approximately 40%)

• Autism spectrum disorders, language disorders, and depression (approximately 15%)

Zuanazzi emphasized that the dataset is “multimodal,” meaning it combines many different forms of data collection, including:

• Questionnaires

• Cognitive testing

• Language assessments

• Emotional and psychological evaluations

• Substance use and addiction data

• Medical status data

• EEG and eye-tracking

• MRI imaging

• Biological samples

• Behavioral monitoring technology

The workshop specifically focused on behavioral monitoring through actigraphy.

Introduction to Actigraphy

Zuanazzi explained that actigraphy is a non-invasive method used to monitor physical activity and sleep continuously over time using wristwatch-like devices equipped with:

• Accelerometers

• Light sensors

• Temperature sensors

Participants in the Healthy Brain Network wore these devices continuously for up to 30 days, producing second-by-second recordings of activity and sleep behavior.

The project used two scientific actigraphy devices:

• Actigraph

• GENEActiv

Zuanazzi contrasted scientific actigraphy watches with commercial smartwatches.

Scientific actigraphy devices provide:

• Raw unprocessed data

• Continuous high-frequency recordings

• No interactive features or apps

Consumer smartwatches instead provide:

• Pre-processed summaries

• Steps

• Heart rate estimates

• Sleep summaries

• Notifications and app functionality

Why Actigraphy Matters for Mental Health Research

Zuanazzi described actigraphy as particularly valuable for mental health research because many conditions — including ADHD, anxiety, and depression — are associated with distinct:

• Physical activity patterns

• Hyperactivity or inactivity

• Sleep timing

• Sleep interruption

• Irregular movement behavior

Actigraphy provides objective real-world measurements rather than relying entirely on self-report questionnaires.

She emphasized several advantages:

• Continuous monitoring over long periods

• Naturalistic recording in homes and schools

• Objective behavioral measurement

• Compatibility with ecological momentary assessment data

• Scalability and relatively low cost

Introduction to the Hands-On Coding Session

Adam Santorelli then led the technical portion of the workshop.

He explained that the hands-on demonstration used data from two Healthy Brain Network participants:

• A 5-year-old girl with ADHD

• A 14-year-old boy with ADHD

To keep processing manageable during the workshop, the data were truncated to one week rather than the full month-long recordings.

Santorelli introduced several key actigraphy concepts:

Tri-Axial Accelerometer Data

The devices record acceleration in three dimensions:

• X-axis

• Y-axis

• Z-axis

measured relative to Earth’s gravity.

ENMO (Euclidean Norm Minus One)

ENMO is a common actigraphy metric measuring physical movement intensity.

Santorelli explained:

• Values near zero indicate little movement or stillness

• Larger values indicate vigorous movement

Angle Z

Angle Z represents the orientation of the wrist relative to the horizontal plane and is heavily used in sleep detection algorithms.

wristpy – Open Source Actigraphy Processing Software

Santorelli introduced wristpy, an open-source Python toolbox developed at the Child Mind Institute for processing raw actigraphy data.

The software:

• Reads raw actigraphy files

• Processes accelerometer recordings

• Extracts movement metrics

• Detects sleep

• Produces analysis-ready CSV outputs

The workshop used Google Colab notebooks for hands-on demonstrations.

Participants installed wristpy using Python’s pip package manager and downloaded example datasets from cloud storage buckets.

Reading and Inspecting Raw Data

Santorelli demonstrated how wristpy loads raw watch data into Python objects.

The raw data included:

• Acceleration measurements

• Luminosity values

• Temperature

• Battery information

• Non-wear indicators

The demonstration revealed that even one week of actigraphy data contains tens of millions of measurements.

Participants then visualized the raw accelerometer traces, observing:

• High-activity daytime periods

• Low-activity nighttime periods likely corresponding to sleep

Running the wristpy Processing Pipeline

Santorelli next demonstrated wristpy’s orchestrator pipeline, which automatically processes raw data into usable behavioral metrics.

The pipeline produced:

• ENMO movement metrics

• Physical activity classifications

• Sleep status predictions

• Angle Z measurements

• Non-wear detection

• Sleep period information

The system also exports:

• CSV result files

• JSON metadata files documenting processing parameters

Santorelli explained that researchers can customize many parameters, including:

• Activity thresholds

• Calibration methods

• Epoch lengths

• Activity metrics

• Non-wear algorithms

Physical Activity Analysis

Using interactive Plotly visualizations, Santorelli compared activity patterns between the two participants.

Findings included:

• The 5-year-old showed higher peaks of vigorous activity

• The 14-year-old remained active later at night

• Younger children generally displayed greater movement intensity

He also demonstrated how researchers can categorize activity into:

• Sedentary

• Light activity

• Moderate activity

• Vigorous activity

based on ENMO thresholds drawn from published literature.

Sleep Detection and Angle Z

Santorelli then demonstrated sleep detection outputs.

Sleep status strongly correlated with reduced fluctuations in Angle Z, reflecting reduced wrist movement during sleep.

The comparison between participants revealed expected developmental differences:

• The 5-year-old generally slept earlier

• The 14-year-old stayed awake later into the night

Challenges in Detecting Non-Wear Versus Sleep

A particularly important methodological discussion involved differentiating:

• Very still sleep

• Device non-wear

Santorelli showed nighttime periods where the participant appeared motionless for extended intervals. These periods could indicate either:

• Deep still sleep

• Removal of the watch before bed

The algorithm currently relies primarily on wrist-angle changes, making it difficult to distinguish perfectly still sleep from actual device removal.

Santorelli suggested that incorporating additional signals such as:

• Temperature

• Capacitive sensors

could improve future algorithms.

NotSleepy – Sleep Analysis Toolbox

Santorelli next introduced NotSleepy, another Child Mind Institute software package designed for extracting sleep metrics from wristpy outputs.

NotSleepy computes nightly sleep measures including:

• Sleep duration

• Time in bed

• Wake after sleep onset (WASO)

• Sleep efficiency

• Number of awakenings

The software also handles daylight savings time adjustments, which Santorelli described as a surprisingly difficult technical problem in sleep analysis.

Computing Daily Physical Activity Metrics

Santorelli demonstrated how processed wristpy outputs can be aggregated into daily activity summaries using the Polars data-processing library.

Metrics included:

• Sedentary duration

• Light activity duration

• Moderate activity duration

• Vigorous activity duration

• Moderate-to-vigorous physical activity (MVPA)

• Total activity counts

The results again showed:

• Higher vigorous activity in the 5-year-old participant

• Greater total daily movement among younger children

Quality Control Challenges in Actigraphy Research

Nathalia Bianchini Esper then shifted the discussion toward quality control and research methodology.

She stressed that actigraphy analysis is not complete after automated processing because:

• Watches may not be worn continuously

• Algorithms can misidentify sleep

• Manual review remains essential

Esper emphasized that most existing actigraphy pipelines still rely partly on:

• Sleep diaries

• Human annotation

• Manual correction

ActiSleep Tracker

To simplify manual quality control, the Child Mind Institute developed another open-source tool called ActiSleep Tracker.

The web-based application allows researchers to:

• Visualize ENMO and Angle Z signals

• Inspect nightly sleep predictions

• Correct sleep windows manually

• Annotate naps

• Adjust wake-up times

• Export corrected annotations automatically

Esper demonstrated how researchers can move sliders interactively to correct algorithm-generated sleep periods.

For example, one participant showed an unrealistically long still period during the night that appeared suspiciously like non-wear rather than sleep. Researchers could manually adjust the wake-up time and automatically generate corrected CSV annotations.

Diagnoses and Comorbidity in the Dataset

Esper reviewed diagnostic overlap within the Healthy Brain Network.

Many participants have multiple diagnoses simultaneously, including combinations such as:

• ADHD and autism

• ADHD and anxiety

• ADHD and learning disorders

She demonstrated how researchers can select subgroups depending on their study goals.

Importance of Seasonality

Esper highlighted seasonality as a frequently overlooked variable in actigraphy research.

Children wearing watches during:

• Winter

• Summer camp periods

• School sessions

may show dramatically different activity patterns unrelated to diagnosis.

Researchers therefore need to consider:

• Enrollment season

• School schedules

• Environmental context

when analyzing physical activity data.

Compliance Analysis

Esper presented compliance analyses showing how many participants meet different thresholds for valid data collection.

Examples included:

• Participants wearing watches continuously for eight days

• Participants wearing watches for only partial days

• Sleep-only compliance thresholds

Less restrictive thresholds naturally increase usable sample sizes.

Sleep and Activity Findings Across Age Groups

Esper showed several aggregate findings from the HBN actigraphy data.

Sleep Timing

Older children and adolescents:

• Tend to go to bed later

• Still wake up at similar weekday times due to school schedules

ENMO Activity Trajectories

Average ENMO curves showed:

• Younger children move more during the day

• Older children show lower activity levels

• Weekend activity patterns differ from weekday patterns

Sedentary and Vigorous Activity

Older children spend:

• More time in sedentary behavior

• Less time in vigorous activity

Girls in the HBN dataset also tended to move less than boys on average.

Open Science and Future Data Release

Esper concluded by emphasizing the broader goals of the project:

• Open science

• Data sharing

• Reproducible research

• Low-cost behavioral monitoring tools

The team announced that the Healthy Brain Network actigraphy dataset is expected to become publicly available during the summer of 2026.

Closing Remarks

Arianna Zuanazzi closed the workshop by thanking Open Data Week organizers, BetaNYC, NYC OTI, and Data Through Design. She reminded participants that all workshop resources — including notebooks, software tools, slides, and repositories — were shared publicly and encouraged attendees to contact the Child Mind Institute team with additional questions.

RESOURCES

• Child Mind Institute — independent nonprofit behind the workshop, focused on children’s mental health, open science, and open data

• Healthy Brain Network (HBN) — community-referred research initiative providing no-cost evaluations and the multimodal open dataset used in the workshop

• HBN Impact — overview of the biobank’s scale and the research it has enabled

• wristpy — CMI’s open-source Python toolbox for processing raw actigraphy data, demoed by Adam Santorelli

• wristpy (JOSS paper) — peer-reviewed Journal of Open Source Software publication describing the wristpy package

• ActiSleep Tracker — CMI web app for manually annotating and quality-controlling actigraphy sleep predictions, demoed by Nathalia Bianchini Esper

• actfast — fast Rust-based actigraphy file reader underlying the wristpy pipeline

• childmindresearch on GitHub — CMI’s open-source repositories, including the actigraphy and sleep-analysis toolboxes

• Polars — fast DataFrame library used in the workshop to compute daily physical-activity metrics

• Plotly for Python — interactive plotting library used to visualize ENMO and sleep traces

VIDEO | AUDIO | RECAP EN / ES / FR | ARCHIVE | PERMALINK

Speakers: Zohran Mamdani - Mayor, New York City; Lisa Gelobter - Commissioner and Chief Technology Officer, Office of Technology and Innovation; Dr. Rajiv Shah - President, The Rockefeller Foundation; Carmen De La Rosa - Council Member, New York City Council; Kristen Gonzalez - New York State Senator, Chair, Internet and Technology Committee.

Overview

New York City announced the creation of five Public Interest Technology (PIT) Crews within the Office of Technology and Innovation (OTI). The multidisciplinary teams will work with city agencies to rapidly design and deploy digital services intended to improve public access to government, beginning with an online portal supporting the city’s new Click-to-Cancel consumer protection rule. The initiative combines city funding with philanthropic support from The Rockefeller Foundation and is intended to demonstrate a faster, user-centered model for delivering digital government services.

Mayor Zohran Mamdani: Applying Product Development to Government

Zohran Mamdani introduced the initiative by drawing an analogy with NASCAR pit crews, describing how highly coordinated teams transformed race strategy through specialization and continuous practice. He said New York City intended to bring the same philosophy to government through Public Interest Technology Crews.

He explained that each PIT Crew would include product managers, designers, software engineers, user researchers, and data specialists working directly with agencies to replace paper-heavy government processes with modern digital services. Rather than waiting years for traditional technology procurements, the teams would move from concept to implementation within months.

The Mayor said the first team would focus on implementing New York City’s newly adopted Click-to-Cancel rule by building an online complaint portal for residents who encounter businesses using difficult or deceptive subscription cancellation practices. The platform would both simplify consumer complaints and provide enforcement data for the Department of Consumer and Worker Protection.

He argued that improving digital government is not simply about efficiency but about rebuilding public trust. Every confusing government website or dead-end process weakens confidence in public institutions, while intuitive digital services demonstrate that government can effectively serve residents.

Mamdani concluded by encouraging experienced technologists to apply to join the PIT Crews, arguing that assembling strong multidisciplinary teams would substantially improve government performance and outcomes for New Yorkers.

Lisa Gelobter: Technology as a Strategic Public Asset

Lisa Gelobter described the launch as the beginning of a new direction for the Office of Technology and Innovation. She thanked the Mayor, Deputy Mayor Julia Kerson, the Rockefeller Foundation, and numerous partners for supporting the initiative.

She said OTI intended to treat technology as a strategic public asset, emphasizing user-first design and measurable outcomes rather than simply delivering software. Her objective, she said, was to transform how New Yorkers interact with government by making digital services simpler, faster, and more effective.

Drawing on experience in both the private sector and federal government, Gelobter said she had consistently worked on projects designed to democratize access and create systemic change. She argued that New York City could now apply those same principles through technology combined with public policy.

She also invited technologists interested in public service to apply for PIT Crew positions, describing the initiative as an opportunity to make a direct impact on millions of residents.

Dr. Rajiv Shah: Philanthropy as a Catalyst

Dr. Rajiv Shah said The Rockefeller Foundation supports initiatives that produce measurable improvements for vulnerable communities and viewed the PIT Crew program as an opportunity to strengthen public service delivery.

He described philanthropy as providing an initial “speed boost” that enables governments willing to innovate to deliver results more rapidly. Rather than celebrating symbolic victories, he said successful public-private partnerships improve everyday lives through better access to affordable housing, tax assistance, consumer protections, and other essential services.

To support the effort, Shah announced an initial commitment of more than $2 million from The Rockefeller Foundation and encouraged other philanthropic organizations to invest in similar collaborations with city government.

Carmen De La Rosa: Innovation as a Public Good

Carmen De La Rosa said the initiative reflected a vision of innovation centered on public benefit rather than technology for its own sake.

She noted that residents regularly seek assistance navigating complex city systems in order to access housing, employment, language services, and other essential government programs. The PIT Crew initiative, she said, represented an effort to close those accessibility gaps.

As Chair of the City Council Technology Committee, she argued that technology should remove barriers rather than create them, while also acknowledging the need to address longstanding inequities that have disproportionately affected many communities. She expressed support for collaboration between the City Council and the administration to modernize government services.

Kristen Gonzalez: Good Technology, Good Government

Kristen Gonzalez said effective technology policy should improve people’s lives, while effective government should make public services easier to access.

She argued that the PIT Crew initiative brought those two principles together by using technology to solve practical problems faced by residents. The first project, she noted, would create a reporting portal allowing consumers to report businesses violating the city’s subscription cancellation rules, strengthening enforcement while returning money to residents.

Speaking as a former technology product manager, Gonzalez encouraged technology professionals to consider careers in public service, arguing that applying technical expertise within government could produce meaningful public benefit.

Questions and Answers

During the media briefing, a reporter questioned whether allocating approximately $5.4 million to the initiative was justified given its focus on improving efficiency.

Zohran Mamdani responded that the first Click-to-Cancel project alone was projected to save New Yorkers more than $160 million annually. He argued that the investment should also be evaluated in terms of reducing frustration and time spent navigating difficult subscription cancellation processes, while noting that Rockefeller Foundation funding further strengthened the initiative.

Asked about implementation timelines, Mamdani said the first Click-to-Cancel product was expected to launch by the fall of 2026.

Lisa Gelobter added that agencies across city government were being invited to propose projects. Selection criteria would prioritize initiatives that directly serve New Yorkers, advance mayoral priorities, and have strong agency commitment. She emphasized that designing services around user needs would both improve outcomes and reduce waste by ensuring technology investments addressed genuine public requirements.

COMMENTARY

• Mamdani invests in tech capacity to “solve real problems” — Don Moynihan on PIT Crew as a post-DOGE progressive model for state capacity

• Decades in the making: New York City launches its PIT Crew! — BetaNYC traces ten years of advocacy for in-house civic tech

• Mamdani takes City Hall tech for a spin with five new ‘PIT crews’ — amNewYork confirms $5.24M baselined, 30 hires, salary bands, no civil service exam

• Mamdani’s new ‘PIT Crew’ tech teams to help New York agencies improve digital services — StateScoop’s govtech read on the launch

• NYC Mayor Launches Tech Teams to Tackle Agency Issues — Government Technology, with Click to Cancel’s Oct. 1 effective date and $525 penalties

• Mamdani forms Public Interest Technology crews to improve NYC government efficiency and affordability — ABC7 New York’s broadcast coverage

• New York City Is Set to Hire a Swarm of Tech Experts — Inc. on the hiring push

RESOURCES

• PIT Crew — job postings for product managers, engineers, designers, and researchers

• Mayor Mamdani Launches “Public Interest Technology (PIT) Crew” — official announcement, July 13, 2026

• Official transcript of the Luna Park announcement — City Hall’s own record of the event

• DCWP release on the PIT Crew and Click to Cancel — the agency partnering on the first project

• The Rockefeller Foundation on the PIT Crew launch — funder’s announcement of its $2M-plus commitment

• Dr. Rajiv Shah — President, The Rockefeller Foundation

• Council Member Carmen De La Rosa — Chair, City Council Committee on Technology

• Senator Kristen Gonzalez — Chair, State Senate Internet and Technology Committee

• BetaNYC — civic tech organization that pressed for in-house city digital capacity

• tEQuitable — the company Lisa Gelobter founded before becoming NYC CTO

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speaker: Ruoyu Li - Graduate Student, NYU Center for Urban Science + Progress (CUSP); Graduate Research Assistant, NYU Light Pollution Project; DarkSky International Advocate
Moderator: Michael Xie - NYC Open Data, NYC OTI

Introduction and Personal Background

Ruoyu Li introduces himself as a graduate student at NYU’s Center for Urban Science and Progress (CUSP), a graduate research assistant on a light pollution project, and an advocate with DarkSky International. He explains that he has worked on light pollution advocacy for more than a decade through presentations, public outreach, research, and community organizing.

He also discusses his involvement with:

• ASPRS NYU Student Chapter

• DarkSky International

• DarkSky New York

• International advocacy efforts in both New York and Beijing

He explains that DarkSky International’s mission is to protect communities and ecosystems from harmful effects of light pollution through:

• Outreach

• Advocacy

• Conservation

• Better lighting design

Recent New York State Light Pollution Legislation

Li begins by discussing recent media attention surrounding proposed New York State legislation on outdoor lighting.

He explains that many headlines inaccurately described the proposal as requiring lights to be turned off after 11 p.m., while the actual bill:

• Allows essential lighting

• Encourages proper shielding

• Supports timers and motion sensors

• Focuses on responsible lighting rather than total darkness

He identifies two companion bills:

• Assembly Bill A4615

• Senate Bill S5007

Participants are encouraged to:

• Check whether their representatives sponsor the bills

• Contact legislators

• Advocate for stronger support

Li emphasizes that public awareness is currently one of the most important aspects of light pollution advocacy.

Global Examples of Light Pollution Regulation

The presentation expands beyond New York to discuss legislation and regulation elsewhere.

Examples include:

• Puerto Rico’s dedicated light pollution law (2008, later revised)

• European environmental regulations

• China’s newly approved Ecological and Environmental Code

Li highlights Puerto Rico as a particularly important case because it combines:

• Legal regulation

• Wildlife protection

• Community advocacy

One example involves sea turtle conservation, where beaches use red lighting to avoid disorienting hatchlings that would otherwise move toward bright roads and urban lighting instead of the ocean.

Another Puerto Rico example describes a local advocate successfully convincing officials to remove harmful lighting near a lighthouse that had previously destroyed dark-sky viewing conditions.

Defining Light Pollution

Using DarkSky International’s definition, Li explains light pollution as:
“human-made alteration of outdoor light levels from those occurring naturally.”

He distinguishes between:

• Necessary lighting for safety and function

• Excessive or wasted lighting

He cites research estimating:

• Light pollution increasing roughly 10% annually

• 80% of the global population living under sky glow

• 99% of people in the U.S. and Europe unable to see the Milky Way

Types of Light Pollution

Li outlines major forms of light pollution:

• Glare

• Light trespass

• Sky glow

Examples include:

• Blinding vehicle headlights

• Billboards

• Neighboring lights shining into homes

• Urban sky glow reflecting off clouds and water

He contrasts heavily light-polluted cities with Flagstaff, Arizona, which adopted early dark-sky regulations because of its astronomical observatory.

Flagstaff is presented as evidence that:

• Economic growth does not require severe light pollution

• Legislation and proper lighting design can significantly reduce sky glow

Additional examples include:

• Fishing boats using powerful lights

• Daytime glare from reflective buildings

• Bird collisions caused by glass facades

Lighting, Safety, and Public Housing

Li challenges the idea that brighter lighting automatically improves safety.

Using visual comparisons, he argues that:

• Poorly designed lighting can blind observers

• Excessive brightness can hide potential threats

• Shielded lighting often improves visibility

He discusses controversial floodlight experiments conducted in New York public housing developments, where researchers claimed brighter lighting reduced crime.

Li argues that:

• The conclusions are debatable

• Residents often disliked the lighting

• The issue raises concerns about environmental equity

He contrasts wealthier communities, where residents often influence lighting decisions, with public housing residents who may have little control over lighting conditions.

Ecological Impacts

A substantial portion of the session focuses on ecological consequences.

Topics include:

• Sea turtle disorientation

• Bird migration disruption

• Insect decline

• Fish and marine ecosystem disruption

• Impacts on bats and nocturnal wildlife

Li explains that New York lies on a major migratory bird corridor and cites estimates that approximately one billion birds annually die from building collisions.

He discusses the “Tribute in Light” memorial in New York City as a well-known example where birds become trapped in powerful light beams. Volunteers now monitor migration conditions and temporarily shut off the lights when large numbers of birds become trapped.

The presentation also discusses effects on:

• Plant growth cycles

• Seasonal timing

• Agricultural systems

Examples include trees and crops showing altered growth patterns near artificial lighting.

Human Health and Circadian Rhythms

Li explains that human biology evolved around natural cycles of darkness and light.

He discusses:

• Melatonin disruption

• Sleep disturbance

• Blue light exposure

• Possible links to disease and cancer risk

Examples include:

• LED lighting with strong blue wavelength peaks

• Phones and computer screens

• Night mode features designed to reduce blue light exposure

He contrasts:

• Bright blue-rich daylight conditions
  with

• Warmer amber lighting more appropriate for nighttime environments

Responsible Lighting Principles

The presentation introduces DarkSky International’s “Five Principles for Responsible Outdoor Lighting.”

Key concepts include:

• Use light only when necessary

• Direct light only where needed

• Use the lowest effective brightness

• Control lighting with timers and sensors

• Prefer warmer color temperatures

Li shows examples of:

• Shielded residential lighting

• Well-designed sports field lighting

• Poorly shielded park and city lighting

He repeatedly emphasizes:
“More light does not mean more safety.”

Equity and the Dark Sky Movement

Li argues that dark sky preservation should not become a privilege only available to wealthier communities.

He discusses international dark sky designations and notes that:

• Many are located in wealthier regions

• Major cities are increasingly showing interest in dark-sky initiatives

A notable example is Shenzhen, China, becoming an International Dark Sky Community despite being a dense megacity.

Li argues that awareness should be incorporated during periods of rapid urbanization so that cities avoid expensive retroactive corrections later.

New York City Examples

The presentation includes several local examples of excessive or poorly managed lighting:

• Parking lot lighting left on overnight

• Reflections from new skyscrapers

• Decorative building lighting

• Floodlighting in parks

He specifically discusses:

• The new JPMorgan building

• Con Edison parking lot lighting

• Lighting in Central Park

Li also reviews existing NYC legislation passed in 2018 regulating:

• Lighting in city-owned buildings

• Bird migration protections during peak migration periods

He notes that many previous state and city bills addressing light pollution have failed or stalled despite repeated reintroduction.

Open Data and Light Pollution Research

The final section focuses on measurement and open data.

Li discusses NYC 311 complaint data related to:

• Parking lot glare

• Excessively bright signs

• Street lighting complaints

He demonstrates how NYC Open Data allows researchers to map complaint locations and identify recurring issues.

He also highlights publicly available tools and datasets including:

• Light Pollution Map

• VIIRS satellite data

• NASA Black Marble

• Google Earth Engine datasets

• Colorado School of Mines nighttime light composites

• DMSP historical satellite imagery

• SDGSAT-1 satellite imagery

Applications discussed include:

• Disaster recovery analysis

• Energy studies

• War monitoring

• Long-term urbanization analysis

Li explains that newer satellites such as SDGSAT-1 provide:

• Higher spatial resolution

• Multiple spectral bands

• Better visibility of blue wavelengths

Citizen Science and Community Participation

Audience discussion includes community science initiatives such as Globe at Night.

Participants can:

• Observe constellations

• Compare visibility conditions

• Submit observations online

Li explains that citizen science datasets have already been used in published research estimating annual increases in global light pollution.

Audience Questions

Questions address:

• Construction lighting

• Air pollution interactions

• Citizen science

• Blue wavelength impacts

• Open datasets desired from NYC agencies

Li says one dataset he would particularly like released publicly is a detailed inventory of NYC streetlights, including:

• Locations

• Fixture types

• Installation dates

• Lighting designs

He notes that DOT declined requests for some of this information due to stated security concerns.

Conclusion

Li concludes by emphasizing that light pollution should be treated as an environmental pollution issue similar to:

• Air pollution

• Noise pollution

He argues that solutions already exist through:

• Better design

• Responsible use

• Regulation

• Public awareness

• Open data

• Community advocacy

The session closes with encouragement for participants to:

• Support legislation

• Join advocacy organizations

• Conduct research

• Promote responsible lighting practices

• Use open data tools to study and visualize light pollution impacts.

RESOURCES

• DarkSky International — the advocacy organization whose mission and dark sky place certification program Ruoyu Li presented

• NYU Center for Urban Science + Progress (CUSP) — the urban data science graduate center where the speaker is a graduate student and research assistant

• Five Principles for Responsible Outdoor Lighting — the DarkSky–IES framework for useful, targeted, low-level, controlled, warm-colored lighting

• DarkSky Outdoor Lighting Code Templates — free legal and technical ordinance templates advocates can send to local representatives

• NASA Black Marble — VIIRS-derived nighttime lights product suite used to study energy, disasters, and light pollution

• Earth Observation Group VIIRS Nighttime Lights — annual composite nighttime light data from the Colorado School of Mines

• Globe at Night — international citizen science program for measuring and submitting night sky brightness observations

• Light Pollution Map — interactive map with layered VIIRS satellite and sky brightness data discussed as a public-friendly visualization tool

• NYC 311 Service Requests dataset — NYC Open Data source for parking lot glare and streetlight complaints from 2020 to present

• NYS Assembly Bill A4615 — proposed New York State light pollution legislation (identical companion to Senate Bill S5007)

Executive Order No. 62 – Establishing a Temporary Moratorium on Data Centers in New York While the State Develops Higher Standards for Data Center Development and a Benefits Blueprint to Support Localities - issued by New York Governor Kathy Hochul at Albany and filed with the New York State Department of State -14 July 2026

Overview

Governor Kathy Hochul issued Executive Order No. 62, pausing New York State discretionary environmental permitting for large data centers while the State builds a regulatory framework for the buildout of AI and cloud computing infrastructure. The order cites unprecedented growth in demand for data center development, driven by computing infrastructure supporting artificial intelligence, cloud computing, streaming, and other computing operations, and raises concerns over electricity consumption, water use and treatment, local environmental impacts, grid reliability, stranded-asset risk, and the fair allocation of infrastructure costs.

The order notes that as of May 2026, nearly 12 gigawatts (12,000 megawatts) of data center load requests sit in the New York Independent System Operator interconnection queue, with more than eight gigawatts entering the queue in 2025 alone.

The order builds on Energize NY Development, announced in the 2026 State of the State, which directs the Public Service Commission to modernize how large energy consumers — including data centers — connect to the grid while ensuring those consumers pay their fair share or supply their own power. It is the policy of the State that the cost of electric system upgrades required to serve large loads should not be paid for by every-day New Yorkers.

Rationale

The order argues that existing State regulatory frameworks are not yet prepared to address:

• Large-scale water use and treatment that could strain aquifers, surface waters, and public infrastructure.

• Growth in statewide electric load that challenges the State’s clean energy targets and requires procurement of additional supply.

• Unpredictable development that creates risk for utilities and ratepayers when infrastructure investments are made in anticipation of loads that may not fully materialize.

• Siting and operation impacts on energy use, water quality, air quality, noise, lighting, and quality of life, which New Yorkers have raised as legitimate concerns.

• Increasing competition for clean freshwater amid water-quality threats and climate-driven changes in precipitation and drought, making water reuse and best available conservation technologies increasingly important in high-water-demand sectors.

It also notes that while negotiating local benefits with developers is the responsibility of the locality, the State can offer technical resources and best practices to support those negotiations.

1. Data Center Permitting Moratorium and Generic Environmental Impact Statement (GEIS)

The Department of Public Service (DPS) is directed to examine the impacts of data center interconnection to the electric distribution network through its proceeding under Case 26-E-0045, Proceeding on Motion of the Commission to Address Interconnection Reforms for Large Loads. In connection with that proceeding, DPS is further directed to initiate a formal public process — including public comment and a public hearing — to create a Generic Environmental Impact Statement under the State Environmental Quality Review Act (Article 8 of the Environmental Conservation Law and its regulations), assessing the potential environmental impacts of data center construction and operation in the State, including energy demand, water use and quality, air quality, disproportionate impacts on disadvantaged communities, and noise levels. DPS is to submit a report of the Final GEIS and findings statement, consulting with the Department of Environmental Conservation (DEC) and other relevant agencies.

Until DPS submits that report:

• DEC is directed to hold in abeyance all applications for any discretionary permit, approval, license, or similar permission for the construction or expansion of a data center that are or may be pending before DEC and that DEC has not determined to be complete before the date of the order.

• As a condition precedent to a determination of completeness, DEC may require the applicant to identify and describe in writing whether the application relates to or involves the construction or operation of a data center.

• The provision does not apply to permits, approvals, licenses, or similar permissions from local governments.

• DEC is to assist DPS in preparing the GEIS.

2. Developing a Community Investment Framework

Empire State Development (ESD) is directed, within 60 days, to consider feedback on and create and post on its website a Community Investment Framework, to assist localities in analyzing and attaining local economic benefits and mitigating potential negative effects of hosting a data center.

The Framework is to include guidance associated with:

• Creation and maintenance of a community investment fund into which developers or operators provide capital usable for energy affordability efforts and enhancements to public services such as child care, K-12 programming, or public infrastructure.

• Investments in local infrastructure such as local energy distribution systems, broadband or irrigation systems, or wastewater treatment plants.

• Frameworks giving organized labor a seat at the table, prioritizing prevailing wage standards and project labor agreements for construction, local hiring, apprenticeships, and workforce development.

• Transparency through reporting requirements or other means, so communities understand key economic metrics associated with data center development.

Localities and other governmental entities — including but not limited to Industrial Development Agencies — may use the Framework to negotiate terms and conditions with developers or operators.

3. The New York Grid Acceleration Fund

DPS is directed to consider the development of a mechanism to protect all customers from the risk of significant costs and stranded-asset risks, including consideration of a New York Grid Acceleration Fund, and may consider such a mechanism as part of the Energize NY Proceeding.

The Fund may require data centers to:

• Make upfront capital contributions to finance grid improvements.

• Participate in demand response programs.

• Support procurement of new clean energy supply, including distributed energy resources.

• Contribute to an insurance pool to which developers may need to contribute.

DPS is to consider how contributions may be structured, including contribution levels and allocation, and to consider developing a process to work with utilities and stakeholders to identify necessary infrastructure improvements. A component may include measures protecting ratepayers from project delays, scope changes, or cancellations resulting in stranded assets. DPS may also evaluate approaches requiring data centers to fund new clean electric generation and/or dedicated battery storage, consistent with the State’s clean energy goals, including customer-sited distributed energy resources, to the greatest extent feasible. The Fund could also explore options to support energy affordability.

4. Interconnection, Reliability, and Cost Allocation

Within 60 days, DPS is directed to form a Data Center Interconnection Working Group to identify and resolve issues related to the interconnection of data centers and other large loads, supporting efficient interconnection of large new customers and faithful compliance with “beneficiary pays” principles as related to network upgrade and resource adequacy costs.

DPS is also directed to convene the State’s transmission owners to review their practices and methodologies for studying the system impacts of data centers and other large loads, in order to understand their sufficiency for estimating and managing cost impacts — both as they relate to network upgrades and to supply. DPS is directed to report to the Commission within 90 days.

Data centers may also be subject to service classifications and requirements developed by DPS and as may be established by the Public Service Commission.

5. Data Center Water Withdrawal Review and Report

DEC is to assess whether new or amended regulations, policies, reporting, or guidance is necessary or appropriate to help ensure its water withdrawal program requirements — pursuant to 6 NYCRR Parts 601 and 602 — accurately and completely reflect the water demands of large use customers in the State, including data centers.

No later than twelve months after the date of the order, DEC is to deliver a report setting forth the results of the assessment and identifying potential regulatory, policy, and guidance actions necessary or appropriate to address concerns associated with the siting and operation of data centers in the State.

6. Definition

For purposes of the order, a “data center” means a facility or group of facilities located on the same site or contiguous sites used to house computer servers, associated components, or computing or telecommunications equipment for the storage, processing, distribution, and/or management of data. Characteristics of covered data centers include servers, associated components, or computing or telecommunications equipment which:

1. Are in facilities containing uninterruptible power supply systems, specialized cooling systems designed for high-density computing loads, and/or contain cybersecurity systems designed for secure digital infrastructure operations;

2. Provide data storage, cloud computing, and/or content delivery to customers, internal operations, and/or affiliated business operations, oftentimes on a continuous twenty-four-hour cycle; and

3. Consume or can consume 50 megawatts of energy or more.

Not covered: facilities primarily used for manufacturing; research (including but not limited to quantum computing research or biomedical research); education (including but not limited to facilities used by accredited New York State colleges and universities to the extent they are engaging in academic research, and the Empire AI consortium, or the institute, as defined in section 361 of the Economic Development Law); or the provision of medical care.

7. Agency Consultation

In implementing the order, DEC, DPS, and ESD are to consult with one another and with additional partner agencies and authorities, including but not limited to:

• Authorities Budget Office.

• Department of Health.

• New York State Energy Research and Development Authority.

• Long Island Power Authority.

• Department of State.

• New York Independent System Operator.

Overall

Executive Order No. 62 pauses DEC discretionary permitting for large data centers while DPS runs a SEQRA Generic Environmental Impact Statement tied to its Large Loads interconnection proceeding, and directs parallel work on community benefits guidance, ratepayer protection against stranded assets, interconnection and cost allocation reform, and water withdrawal regulation. The order seeks to ensure that AI-driven computing infrastructure is developed in a way that protects electric ratepayers, water resources, environmental quality, and host communities while maintaining reliability and advancing the State’s clean energy objectives.



RESOURCES

• Executive Order No. 62 — the full text, filed 14 July 2026

• Case 26-E-0045 — the Large Loads interconnection proceeding the GEIS attaches to

• 2026 State of the State — where Energize NY Development was announced

• Department of Public Service — leads the GEIS, the Grid Acceleration Fund review, and the interconnection working group

• Department of Environmental Conservation — holds discretionary permits in abeyance and runs the water withdrawal review

• Empire State Development — must post the Community Investment Framework within 60 days

• State Environmental Quality Review Act (SEQR) — the Article 8 framework governing the Generic Environmental Impact Statement

• Water Withdrawal Permits and Reporting — the 6 NYCRR Parts 601 and 602 program under review

• New York Independent System Operator — operator of the interconnection queue holding nearly 12 GW of data center load requests

• Empire AI — the consortium excluded from the order’s data center definition

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speaker: Ethel Khanis - High School Chemistry Teacher, Virtual Innovators Academy
Moderator: Aryanna Holder - NYC Open Data

Introduction and Classroom Context

Ethel Khanis introduces the session by explaining that she teaches chemistry, Socratic seminar, and advisory classes at the Virtual Innovators Academy, a fully virtual NYC public high school serving students across all five boroughs. She explains that many students arrive with limited spreadsheet experience and varying levels of computer literacy, despite assumptions that younger students are automatically technologically fluent.

The workshop is designed to show how Google Sheets and NYC Open Data can be used together to teach:

• Spreadsheet skills

• Data literacy

• Critical thinking

• Quantitative reasoning

• Student-driven inquiry

She explains that the class sequence begins with students examining misleading graphs and biased visualizations before moving into survey creation, spreadsheet analysis, and eventually work with NYC Open Data datasets.

Collaborative Google Sheets Exercise

Participants are invited into a shared Google Sheet modeled directly on Khanis’s classroom workflow. The session focuses on hands-on experimentation rather than lecture-style instruction.

Students and participants are asked to:

• Claim workspaces in the sheet

• Practice spreadsheet navigation

• Experiment with formulas

• Troubleshoot collaboratively

Khanis emphasizes that the goal is not simply arriving at the correct answer, but understanding how spreadsheet logic works.

She repeatedly encourages participants to:

• Take risks

• Experiment

• Learn through mistakes

• Help one another solve problems

Teaching Spreadsheet Fundamentals

The workshop covers foundational spreadsheet concepts, including:

• Rows and columns

• Cell references

• Selecting ranges

• Basic formatting

• Spreadsheet tabs

• Sorting

• Simple formulas

Khanis notes that many students struggle with actions often assumed to be basic, especially:

• Click-and-drag selection

• Double-click behavior

• Maintaining cell selection

This becomes part of a broader discussion about digital literacy gaps among students.

Participants discuss how many students are comfortable with phones and apps but less experienced with productivity software and structured computer workflows.

Introducing Spreadsheet Functions

Khanis frames spreadsheet formulas as a form of coding, which she says helps students engage more positively with the material.

Functions introduced include:

• SUM

• AVERAGE

• MODE

• SORT

• VLOOKUP

She explains the logic behind formulas:

• The equals sign begins a command

• Parentheses define inputs

• Functions automate repetitive calculations

Participants practice entering formulas directly into the spreadsheet and observe how results change dynamically.

The workshop repeatedly stresses that spreadsheet tools:

• Reduce repetitive work

• Minimize arithmetic errors

• Allow users to focus on interpretation and analysis

VLOOKUP Demonstration

A major portion of the session focuses on VLOOKUP functionality.

Khanis explains the function conceptually as:

• Finding a value in one location and returning related information from another column

Participants discuss:

• Lookup ranges

• Column indexes

• Search values

• Structured data organization

Brian Levine contributes clarification that the lookup value must appear in the first column of the selected lookup range.

The exercise demonstrates how spreadsheet tools can scale from small classroom exercises to much larger datasets.

Virtual Teaching Challenges

Khanis reflects on the challenges of teaching spreadsheet skills in a virtual environment:

• Students switching between Zoom and Sheets

• Limited screen space

• Difficulty monitoring work in real time

• Troubleshooting technical issues remotely

She explains that collaborative spreadsheets help because she can:

• Watch formulas appear live

• Identify mistakes immediately

• Guide students through corrections

Participants discuss how peer support in the classroom often becomes essential for troubleshooting.

Student Engagement and Learning Outcomes

Khanis reports that spreadsheet activities generated unusually high engagement levels among students.

Examples include:

• Students voluntarily staying after class

• Students helping peers debug formulas

• Active use of class chat for support

• Increased willingness to experiment and problem-solve

She describes how students gradually developed confidence working with data and spreadsheets through repeated collaborative exercises.

The session emphasizes that spreadsheet instruction became not just technical training, but also:

• Confidence-building

• Collaborative learning

• Inquiry-based analysis

Connecting Spreadsheet Skills to Open Data

The workshop positions NYC Open Data as the next stage of student learning.

Potential student projects include analysis of:

• Lead exposure data

• Special education datasets

• Other public-interest datasets relevant to students’ communities

Khanis explains that students eventually move toward:

• Selecting their own datasets

• Conducting independent analysis

• Presenting findings based on public data

The broader educational objective is helping students become active interpreters of real-world information rather than passive consumers of charts and statistics.

Audience Discussion

Audience discussion focuses on several recurring themes:

• Spreadsheet literacy as a core modern skill

• Challenges teaching technical concepts remotely

• The role of peer learning

• Assumptions about “digital natives”

• The growing role of AI tools in spreadsheet work

Participants discuss whether students should first learn spreadsheet logic manually before relying on AI-assisted tools such as Gemini.

Khanis argues that conceptual understanding remains essential even when automation tools are available.

Conclusion

The session demonstrates how spreadsheet instruction can be integrated with civic and public-interest datasets to create highly interactive, inquiry-driven learning experiences.

By combining:

• Google Sheets

• Collaborative exercises

• Student-created surveys

• NYC Open Data

the workshop illustrates how students can develop both technical and analytical skills while engaging directly with real-world public data.

RESOURCES

• Teaching Google Sheets Functionality to High School Students through Open Data — the NYC Open Data Week 2026 event page for this workshop

• NYC Open Data — the City’s free public data portal Ethel’s students draw datasets from to choose and analyze their own projects

• Children Under 6 yrs with Elevated Blood Lead Levels (BLL) — the elevated-lead dataset Ethel downloaded as a more complex analysis option for students

• VLOOKUP — Google Sheets function reference — official help page for the vertical lookup function demonstrated in the workshop

• XLOOKUP — Google Sheets function reference — official help page for the lookup function raised during the Q&A

• Google Sheets function list — the full catalog of Sheets functions, a reference for students exploring beyond average and lookup

• Learn about NYC Open Data — free virtual classes — 90-minute Open Data Ambassador classes for anyone wanting to get started with City data

• NYC Open Data Week (YouTube) — channel hosting recordings of this and past Open Data Week sessions

Digital Government Outlook 2026: From Foundations to Transformational Impact

Organisation for Economic Co-operation and Development (OECD), 15 June2026.

Overview

The OECD’s first “Digital Government Outlook” provides a comprehensive assessment of digital government maturity across 36 OECD member states and 8 accession candidate countries. Drawing on the 2025 OECD Digital Government Index (DGI) and the Open, Useful and Re-usable Data (OURdata) Index, the report evaluates how governments are progressing from building digital foundations toward delivering measurable public-sector transformation.

The report argues that governments are under pressure from economic volatility, demographic change, geopolitical instability, environmental challenges, and rapidly evolving technologies such as artificial intelligence (AI). While digital technologies and data are now essential infrastructure for modern governance, many public institutions remain constrained by rigid processes, fragmented systems, workforce shortages, and outdated governance models.

The OECD finds that governments have made substantial progress in creating enabling digital infrastructure and governance frameworks, but implementation and operational delivery continue to lag behind strategic ambition. The challenge has shifted from establishing digital foundations to embedding digital transformation throughout the daily operations of government.

Key Findings

The 2025 OECD Digital Government Index rose from 0.61 in 2023 to 0.70 in 2025, representing a 14% improvement across participating countries. The OURdata Index increased from 0.48 to 0.53. However, accession countries continue to trail OECD members, and implementation remains inconsistent across jurisdictions.

The report emphasizes that most countries are stronger at creating digital strategies, legal frameworks, and policy tools than at operational implementation and performance monitoring. Governments continue to struggle to translate policy intent into consistent delivery at scale.

Public trust also remains fragile. Only 52% of people across 30 OECD countries trust their government to use personal data for legitimate purposes, highlighting that digital transformation depends not only on technology but also on governance, transparency, workforce capability, and enforceable safeguards.

OECD Digital Government Policy Framework

The report structures its analysis around the OECD Digital Government Policy Framework, which identifies six dimensions of digital maturity:

• Digital by design

• Data-driven public sector

• Government as a platform

• Open by default

• User-driven government

• Proactiveness

These dimensions are evaluated through four stages of the policy cycle:

• Strategic approach

• Policy levers

• Implementation

• Monitoring

The framework is designed to measure not simply whether governments have digital policies or online services, but whether they possess the institutional capability to implement them coherently across government operations.

Priority Area 1: Strengthening Digital Public Infrastructure and Data Governance

The report finds widespread adoption of foundational digital public infrastructure (DPI), including:

• Digital identity systems

• Single digital gateways

• Data-sharing platforms

• Cloud infrastructure

• Digital notification systems

However, interoperability and operational data sharing remain uneven. Only 63% of public institutions across OECD countries are sharing data through national interoperability systems. Privacy and cybersecurity frameworks are well established, but practical data-management capabilities such as interoperability, quality assurance, and trusted reuse lag behind.

The report highlights the importance of digital identity adoption and trusted data exchange in enabling proactive and joined-up public services. It also notes growing interest in open-source software and cloud technologies to improve resilience, sustainability, and cross-border collaboration.

Priority Area 2: Governing Digital Investment and Building Skills

Most OECD countries now conduct upfront assessments of digital projects, provide dedicated digital-transformation funding, and offer procurement guidance. Nevertheless, governments remain weak at evaluating outcomes after projects are completed. Only one-quarter of countries systematically assess whether digital investments achieved intended results.

The report advocates more iterative and flexible funding approaches that allow governments to experiment, learn, and scale successful projects incrementally instead of relying on rigid multi-year procurement cycles.

Workforce capability is identified as a major bottleneck. Only six OECD countries have dedicated strategies for digital skills in the civil service. Governments increasingly depend on external suppliers for technical expertise, risking long-term erosion of internal capability and institutional knowledge.

The report calls for multidisciplinary teams, continuous professional development, and stronger public-sector digital talent pipelines to support sustainable transformation.

Priority Area 3: Scaling Trustworthy AI in Government

AI adoption is accelerating across OECD governments. Most countries now have AI strategies, oversight institutions, and workforce training initiatives. AI is most commonly used in internal administrative processes and public-service delivery rather than policymaking or accountability functions.

Despite progress, governance mechanisms remain immature:

• Transparency measures are inconsistent.

• Algorithm registers remain limited.

• Procurement guidance for AI exists in only slightly more than half of OECD countries.

• Few governments measure the actual impact of AI deployments.

Only 28% of countries systematically evaluate the results of AI use in government.

The OECD stresses the importance of enforceable controls, transparency mechanisms, practical procurement guidance, and targeted workforce training to support trustworthy and accountable AI adoption. The report also explores emerging concepts such as “agentic AI” systems capable of autonomous decision-support and workflow execution.

Priority Area 4: Delivering Human-Centred and Proactive Services

Most OECD countries now have government-wide service standards and increasingly involve users in service design. However, implementation remains inconsistent, and user feedback is not systematically integrated into continuous improvement processes.

Only 28% of countries systematically measure the burden imposed on users by public services. Governments continue to struggle with fragmented service delivery across agencies and channels.

The report strongly promotes:

• Omni-channel service delivery

• “Once-only” principles for data collection

• Proactive service delivery

• Joined-up service journeys

• Data-driven anticipation of citizen needs

The OECD argues that governments should move from reactive service provision toward anticipatory and proactive models that reduce administrative burdens and simplify interactions with the state.

Broader Themes

Across all chapters, the report repeatedly returns to several broader themes:

• Digital transformation is fundamentally institutional, not merely technological.

• Shared infrastructure and interoperability are essential to resilience and efficiency.

• Trust, transparency, and accountability are prerequisites for successful digital government.

• Workforce capability is as important as technical infrastructure.

• Governments must move beyond siloed modernization projects toward integrated operating models.

The report also emphasizes cross-border collaboration, especially around digital identity, interoperability standards, open-source software, and shared governance frameworks such as the EU’s eIDAS 2.0 and European Digital Identity Wallet initiatives.

Conclusion

The OECD concludes that governments have largely succeeded in establishing the strategic and infrastructural foundations for digital government. The next challenge is operational: embedding digital, data, and AI capabilities into the core machinery of government in ways that measurably improve outcomes for people and businesses.

The report frames this transition as moving from “digital maturity” to “transformational impact.” Governments that succeed will be those able to integrate governance, technology, workforce capability, data stewardship, and human-centred design into coherent, adaptive, and trustworthy public institutions.

RESOURCES

• Digital Government Outlook 2026: From Foundations to Transformational Impact — the full OECD report (OECD Publishing, Paris, 2026)

• Digital Government Outlook 2026 — full report (web) — chapter-by-chapter HTML edition

• Digital Government Index and OURdata Index: 2025 Results and Key Findings — Working Paper No. 90, the index data the Outlook draws on (Feb 2026)

• OECD Digital Government — the programme behind the Outlook and the Digital Government Policy Framework

• Adopting and Governing AI in Government — Outlook chapter on scaling trustworthy AI, including agentic AI

• Strengthening Digital Public Infrastructure and Data Governance — Outlook chapter on DPI, interoperability and open data

• European Digital Identity (EUDI) Regulation — the eIDAS 2.0 framework the report cites as a cross-border model

• EU Digital Identity Wallet — the European Commission wallet initiative referenced in the Outlook

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Hong Yuan - Open Data Administrator, NYC DOT; Esteban Doyle - Senior Planning Coordinator, NYC DOT; Lisa Mae Fiedler - Open Data Manager, MTA; Jaspreet Lal - Graduate Intern, Open Data, MTA; Alec Bardey - Mobility Data Program Lead, NYC DOT; Mark Seaman - Senior Economist, Policy Unit, NYC DOT
Moderator: Hong Yuan - NYC DOT

Introduction and Overview of Mobility Data Collaboration

Hong Yuan opened the session by introducing the collaboration between NYC DOT and the MTA around mobility-related open data. She explained that the presentation would cover multiple transportation datasets spanning pedestrian activity, buses, subways, traffic counting, congestion pricing, and the newly released Citywide Mobility Survey.

Hong emphasized the broader goal of encouraging agencies to work together by:

• Sharing transportation datasets

• Identifying relationships between datasets

• Supporting cross-agency analysis

• Improving public understanding of mobility patterns

She also encouraged participants to continue conversations during an Open Data Week networking event later that evening.

Pedestrian Counts and the Public Realm

Esteban Doyle introduced NYC DOT’s Public Realm Unit, which focuses on improving walkability and pedestrian comfort throughout New York City. He argued that pedestrian activity is historically undercounted compared to vehicle traffic, despite its importance for urban planning and street design.

Doyle stressed that understanding where and how people walk is essential when balancing competing uses of public street space, including:

• Cars

• Buses

• Bicycles

• Pedestrians

• Curb uses

• Deliveries

NYC DOT’s Biannual Pedestrian Count Program

Doyle explained that NYC DOT’s pedestrian count program began in 2007 as a way to track economic activity during the Great Recession. The city now conducts pedestrian counts at 114 locations across all five boroughs.

Count locations include:

• Commercial corridors

• East River bridges

• Harlem River bridges

Counts are collected:

• Twice annually

• During weekday morning peaks

• During weekday evening peaks

• During Saturday midday periods

DOT uses these counts to calculate a “pedestrian volume index,” which compares average pedestrian activity against the original 2007 baseline.

The data showed:

• Significant pedestrian growth during the early 2010s

• Sharp declines during the COVID-19 pandemic

• Partial post-pandemic recovery

• Volumes still below pre-pandemic levels

Doyle cautioned that the dataset reflects only part of the city because many locations are concentrated in Manhattan commercial areas rather than residential neighborhoods.

Limitations of Traditional Pedestrian Counting

Doyle outlined several limitations in the existing pedestrian count program.

Challenges include:

• Limited geographic coverage

• Heavy concentration in Manhattan

• Lack of residential street counts

• Seasonal variability

• Weather-related variability

• Labor-intensive manual counting

The process currently requires:

• Setting up cameras

• Recording footage

• Manual review and counting

• Cataloging results

DOT is therefore exploring alternative approaches for measuring pedestrian activity more comprehensively and efficiently.

The Pedestrian Demand Map

Doyle introduced DOT’s Pedestrian Demand Map, which was created as part of the city’s pedestrian mobility planning process.

Because the city cannot directly count pedestrians everywhere, DOT instead modeled pedestrian demand using “pedestrian generators,” including:

• Parks

• Schools

• Tourist attractions

• Commercial districts

Using these variables, every street in New York City was categorized into one of five demand levels:

• Global

• Regional

• Neighborhood

• Local

• Baseline

These classifications help guide decisions about:

• Sidewalk widths

• Street redesigns

• Pedestrian improvements

• Public realm investments

Future Pedestrian Modeling Efforts

Doyle discussed future ambitions to move beyond static demand mapping toward dynamic pedestrian route modeling.

Potential future analyses include:

• Which side of a street pedestrians choose

• Preferred crosswalks

• Route selection behavior

• Street-level pedestrian flow modeling

MIT recently released a preliminary pedestrian model for New York City, and DOT provided pedestrian count data to help calibrate the system.

DOT expressed strong interest in integrating such modeling techniques into future planning tools.

MTA Bus Route Segment Speed Data

Lisa Mae Fiedler introduced several MTA open datasets hosted on the New York State Open Data Portal.

One major dataset was the Bus Route Segment Speeds dataset, which provides operational speed information for buses between major route stops known as “time points.”

The dataset includes:

• Average speed between time points

• Average travel time

• Distance traveled

• Number of bus trips

aggregated by:

• Month

• Day of week

• Hour of day

The dataset is generated using GPS “Bus Time” pings from more than 6,000 buses, with location updates every 30 seconds.

What Bus Speed Data Captures

Fiedler emphasized that the dataset intentionally reflects the real rider experience.

Speed calculations include:

• Passenger boarding delays

• Stoplights

• Operator changes

• Traffic congestion

• Double-parked vehicles

• Construction slowdowns

She demonstrated how combining segment-speed data with geospatial route and stop datasets allows analysts to build detailed route performance maps.

Examples included visualizations of the M1 bus in Manhattan, showing highly variable speeds across different route segments.

Subway Hourly Ridership Data

Jaspreet Lal presented MTA subway datasets.

The Subway Hourly Ridership dataset provides:

• Hourly ridership estimates

• Station-level entries

• Payment method information

• OMNY versus MetroCard usage

The dataset is updated weekly and is based on turnstile entries.

Visualizations showed:

• January 2024 as the point where OMNY and MetroCard usage converged

• Rapid subsequent dominance of OMNY

• Sharp decline in MetroCard usage

Lal explained that three-month rolling averages were used to smooth fluctuations caused by:

• Weekday seasonality

• Holidays

• Weather

Subway Schedule Data and CBTC

Lal also discussed the Subway Schedules dataset, which includes:

• Base schedules

• Supplemental schedules

• Construction-related service changes

• Holiday schedules

Using the data, the team visualized train arrival frequency distributions during weekday rush hours.

The presentation highlighted the superior consistency of the:

• 7 line

• L line

which use Communication-Based Train Control (CBTC).

CBTC allows trains to:

• Continuously communicate positions

• Operate closer together safely

• Adjust dynamically in real time

resulting in:

• Shorter waits

• More reliable service

Modernizing Vehicle Classification Counts with Computer Vision

Alec Bardey introduced NYC DOT’s effort to modernize vehicle classification counting using computer vision.

Current traffic counting methods rely heavily on:

• Contractors setting up cameras

• Manual review of footage

• Human classification of vehicles

Approximately:

• 80% of costs are associated with manual counting

• Only 20% involve recording and storage

DOT possesses:

• 15 years of traffic video footage

• Corresponding vehicle count data

creating a major opportunity for machine learning applications.

YOLO Computer Vision Prototype

Bardey described the prototype system built using a YOLO computer vision model led by NYMTC fellow and PhD candidate Boshra Khalili.

The system:

• Detects vehicles in video

• Draws bounding boxes

• Tracks movement across threshold lines

• Automatically classifies vehicles

DOT demonstrated a prototype video where vehicles crossing a roadway were automatically identified and counted in real time.

Cleaning 15 Years of Legacy Traffic Data

A major challenge involves standardizing legacy vehicle classification count files.

Bardey showed examples of older spreadsheet formats containing:

• Merged cells

• Inconsistent headers

• Contractor-specific layouts

• Poor formatting

DOT is building automated workflows to:

• Identify file formats

• Clean data

• Standardize structures

• Create centralized databases

The goal is eventually to publish standardized traffic count datasets on NYC Open Data.

Future Computer Vision Applications

DOT hopes the new system will eventually:

• Reduce counting costs

• Expand traffic count coverage

• Enable analysis-ready datasets

• Generate counts from archived footage

Importantly, archived footage originally collected for one purpose — such as truck counts — could later be reused to generate:

• Pedestrian counts

• Bicycle counts

• Additional mobility analyses

Congestion Pricing Open Data

Fiedler next discussed congestion pricing data.

The MTA’s Congestion Relief Zone Vehicle Entries dataset tracks:

• Vehicle crossings into Manhattan south of 60th Street

• Entry locations

• Vehicle classes

• 10-minute intervals

The dataset began with the launch of congestion pricing in January 2025 and updates weekly.

Visualizations showed clear behavioral shifts:

• Drivers rushing into Manhattan just before toll activation at 5 a.m.

• Sharp immediate drops in entries after tolling begins

Fiedler emphasized that open data has been central to public reporting around congestion pricing.

Bridge and Tunnel Crossings Data

Fiedler also introduced MTA Bridges and Tunnels hourly crossing data, which includes:

• Facility-level crossings

• Directional flows

• Vehicle class

• Payment method estimates

The dataset uses:

• Electronic tolling systems

• Automated vehicle classification technology

to estimate:

• E-ZPass usage

• Toll-by-mail usage

• Traffic flows

Fiedler described the dataset as one of the richest available sources for understanding regional traffic movement patterns.

The Citywide Mobility Survey

Mark Seaman presented NYC DOT’s Citywide Mobility Survey (CMS), a household travel survey conducted every two to three years.

The survey captures detailed information about:

• How New Yorkers travel

• Trip purposes

• Transportation modes

• Demographic differences

• Equity impacts

The 2024 survey included:

• 3,500 participants

• More than 100,000 trips

• GPS-tracked travel data

• Smartphone app participation

• Web and phone survey options

Mobility Trends Revealed by CMS

Seaman reviewed several high-level findings.

Mode Share

Walking and driving were approximately tied as the city’s most common transportation modes in 2024.

Other findings included:

• Transit usage still below 2019 levels

• Partial post-pandemic recovery

• Modest but notable increases in biking

Growth in Deliveries

The survey found:

• 41% of NYC households received deliveries on a typical day in 2024

• Up from 31% in 2019

Growth occurred across categories including:

• Packages

• Restaurant takeout

• Grocery deliveries

Open Streets Awareness

CMS also measured public familiarity with Open Streets programs.

Results showed:

• Highest awareness in Manhattan core neighborhoods

• Lower awareness in outer Brooklyn and Queens

Parking Analysis

Using CMS data, DOT analyzed where residents park their cars.

The survey identified areas with especially high levels of on-street parking dependence, including:

• Inner Brooklyn

• Upper Manhattan

• Southern Bronx

This information helps DOT evaluate:

• Parking demand

• EV charging needs

• Potential cruising-for-parking behavior

Estimating Mode Shift to Biking

CMS was also used to estimate what transportation modes bike trips may have replaced.

By comparing bike trip distances to similar trips by other modes, DOT estimated:

• Approximately 36% of citywide bike trips may replace car trips

• Approximately 15% in Manhattan specifically

The analysis helped DOT estimate potential greenhouse gas impacts of cycling growth.

Pedestrian Route Choice Modeling

Seaman described an advanced pedestrian route-choice model built using 17,000 walk trips from the 2019 CMS.

Using GPS traces and statistical modeling, researchers evaluated how pedestrians value street characteristics including:

• Sidewalk width

• Street trees

• Traffic volume

• Street lighting

• Land use

• Crime levels

• Scaffolding

The project became the basis of a doctoral dissertation and is expected to be formally published.

Discussion on Post-Pandemic Pedestrian Declines

During Q&A, participants asked why pedestrian volumes have struggled to fully recover since COVID-19.

Doyle and Seaman pointed to several factors:

• Remote work reducing commuting

• Lower transit ridership

• Growth in deliveries

• Declines in shopping trips

• Increased online shopping

Seaman noted that while total walking trips declined, walking’s share of overall trips remained relatively stable.

Closing Remarks

Hong Yuan concluded the session by encouraging participants to explore the many mobility datasets discussed during the presentation.

She emphasized that NYC Open Data now hosts thousands of datasets and highlighted the importance of continuing cross-agency collaboration to improve mobility analysis and public understanding of transportation trends in New York City.

RESOURCES

• Open Data About Mobility — the NYC Open Data Week session page for this MTA + NYC DOT joint presentation

• MTA Bus Route Segment Speeds: Beginning 2025 — dataset of bus speeds between timepoints, highlighted by Lisa Mae Fiedler

• It’s 2 a.m. Do you know where your bus is? — MTA blog post on the bus-matching algorithm, written by Gayan Seneviratna

• Mapping Movement: Exploring NYC Bus Route Shapes — MTA blog post pairing bus geometries with segment speed data

• MTA Subway Hourly Ridership: Beginning 2025 — hourly ridership by station complex and fare payment class, presented by Jaspreet Lal

• MTA Congestion Relief Zone Vehicle Entries: Beginning 2025 — vehicle crossings into the CRZ in 10-minute intervals

• The Most Detailed View of NYC Traffic (So Far) — MTA blog post explaining the Congestion Relief Zone entries dataset

• MTA Bridges and Tunnels Hourly Crossings: Beginning 2019 — hourly crossings by facility, direction, and vehicle class, created by Niki Keramat

• NYC DOT Citywide Mobility Survey — DOT’s household travel survey, presented by Mark Seaman; 2024 data newly released

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Joli Golden - Data Dissemination Specialist, U.S. Census Bureau; Brian Glassman - Poverty Statistics Branch Chief, U.S. Census Bureau; Monica Dukes - Data Dissemination Specialist, U.S. Census Bureau
Moderator: Joli Golden - U.S. Census Bureau

Introduction to the Census Bureau and the Session Focus

Joli Golden opened the session by introducing herself as a Data Dissemination Specialist with the U.S. Census Bureau. She explained that her work involves helping a wide range of users — including libraries, nonprofits, students, congressional offices, city agencies, and state agencies — understand and access Census data.

Golden described the presentation as a “deep dive into poverty,” intended to help participants understand:

• The official poverty measure

• The supplemental poverty measure (SPM)

• Key Census Bureau poverty datasets

• Public data tools for accessing poverty statistics

She introduced Monica Dukes, another Census Data Dissemination Specialist, and Brian Glassman, Chief of the Census Bureau’s Poverty Statistics Branch, both of whom assisted with audience questions during the session.

Golden also briefly reviewed the Census Bureau’s role as the largest federal statistical agency, highlighting its responsibility for conducting:

• The decennial census

• The American Community Survey (ACS)

• The Economic Census

She emphasized that the Census Bureau often collects data later used and released by other federal agencies, including labor and economic statistics.

Defining Poverty – Official vs. Supplemental Poverty Measures

Golden explained that the Census Bureau releases two major poverty measures each year:

• Official Poverty Measure (OPM)

• Supplemental Poverty Measure (SPM)

Both measures evaluate whether people have enough resources to meet basic needs, but they define resources, family units, and needs very differently.

Official Poverty Measure (OPM)

The official poverty measure:

• Assumes resources are shared only among people related by birth, marriage, or adoption

• Uses pre-tax cash income only

• Relies on a threshold derived from a 1963 minimum food diet multiplied by three and adjusted for inflation

Resources counted include:

• Wages and earnings

• Social Security

• Unemployment income

• Retirement income

• Interest and dividends

• Public assistance

The measure does not include:

• SNAP benefits

• Medicaid

• Housing subsidies

• Tax credits

• Stimulus payments

Golden stressed that the official poverty threshold is geographically uniform. For example:

• A family of two adults and two children had an official poverty threshold of $31,812 in 2024

• The same threshold applied in both Mississippi and New York City

Supplemental Poverty Measure (SPM)

The supplemental poverty measure differs substantially.

SPM:

• Expands the definition of resource-sharing units to include some unrelated household members such as unmarried partners

• Accounts for geographic differences in living costs

• Includes non-cash benefits and tax credits

• Subtracts major expenses from available resources

Resources added include:

• SNAP

• WIC

• Housing subsidies

• School lunch benefits

• Utility assistance

• Tax credits

Resources subtracted include:

• Childcare expenses

• Taxes paid

• Medical expenses

• Child support payments

Golden explained that SPM thresholds vary by housing tenure and geography, producing much higher poverty thresholds in expensive metropolitan regions like New York City.

Household Composition Example

Golden used a hypothetical household to demonstrate differences between the two poverty measures.

The example household included:

• An unmarried couple

• A child

• A grandmother

• An unrelated roommate

Under the official poverty measure:

• The unmarried partner was treated separately

• The unrelated roommate was treated separately

Under the supplemental poverty measure:

• The unmarried couple and related household members were grouped together

• Only the unrelated roommate remained separate

Golden used this example to illustrate how family definitions can significantly affect poverty calculations.

Geographic Differences in Supplemental Poverty Thresholds

Golden demonstrated how SPM thresholds differ geographically by using Census spreadsheets containing metro-area supplemental poverty thresholds.

She showed comparisons between:

• National supplemental poverty thresholds

• New York Metropolitan Statistical Area thresholds

The New York metro region included:

• New York City

• Newark

• Jersey City

• Surrounding suburban counties

The presentation showed that:

• Renters in New York faced the highest supplemental poverty thresholds

• Owners without mortgages faced lower thresholds

• National thresholds remained substantially below New York regional thresholds

Golden emphasized that SPM attempts to reflect real-world cost differences across housing markets.

Akron, Ohio Case Study

Golden presented a detailed example of a family of four renting an apartment in Akron, Ohio in 2023.

The family had:

• $35,000 in pre-tax income

Under the official poverty measure:

• The family was not considered poor because income exceeded the official threshold of $30,900

Under the supplemental poverty measure:

• SNAP benefits and taxes were incorporated

• Childcare and work expenses were deducted

• Effective resources fell to $25,700

• The family was considered poor because resources fell below the supplemental threshold of $33,763

Golden stressed that although such differences receive attention, most households are classified consistently under both systems.

Appropriate Uses of Each Poverty Measure

Golden explained that each poverty measure serves different analytical purposes.

Official Poverty Measure Uses

The official measure is best for:

• Long-term national trend analysis

• Federal grant eligibility

• Standardized national comparisons

because official poverty data extend back to 1959.

Supplemental Poverty Measure Uses

The supplemental measure is better suited for:

• Evaluating government assistance programs

• Examining geographic cost differences

• Measuring effects of tax credits and benefits

• Comparing metropolitan areas and regions

SPM data are available beginning in 2009.

National Poverty Trends and Pandemic Effects

Golden reviewed national poverty statistics for 2024.

According to the Census Bureau:

• Official poverty rate: 10.6%

• Supplemental poverty rate: 12.9%

Charts displayed long-term trends for both measures.

Golden highlighted the dramatic temporary decline in supplemental poverty rates during the COVID-era stimulus period.

As stimulus payments and expanded tax credits boosted household resources:

• Supplemental poverty rates dropped sharply

After many pandemic-era supports expired:

• Supplemental poverty rates increased again

The supplemental poverty measure consistently remained above the official poverty measure throughout most of the period shown.

State-Level Differences Between OPM and SPM

Brian Glassman joined the discussion to explain a state-level map comparing official and supplemental poverty rates.

He explained that some states showed:

• Higher SPM rates than official rates

• Lower SPM rates than official rates

• No statistically significant difference

Glassman stressed that all poverty estimates include statistical uncertainty because they are survey-based estimates.

He also mentioned an upcoming Census Bureau working paper analyzing households classified as poor under one measure but not the other, examining demographic differences between those groups.

Major Census Poverty Datasets

Golden then shifted into a detailed overview of the major Census Bureau datasets used to measure poverty.

Current Population Survey Annual Social and Economic Supplement (CPS ASEC)

Golden described the CPS ASEC as:

• The official source of national poverty estimates

• The longest-running Census poverty survey

• Available back to 1959

The survey:

• Samples approximately 90,000 addresses annually

• Is conducted during February–April

• Supports national one-year estimates and state-level three-year averages

The Census Bureau’s annual poverty reports are based on this dataset.

Golden highlighted an America Counts article demonstrating how educational attainment strongly correlates with higher earnings and lower poverty rates using CPS ASEC data.

She also introduced the Census Bureau’s Microdata Access Tool (MDAT), which allows advanced users to directly analyze CPS microdata.

American Community Survey (ACS)

Golden described the ACS as the “gold standard” of Census Bureau data.

The ACS:

• Samples approximately 3.5 million addresses annually

• Collects roughly 2.2 million responses

• Produces detailed demographic, housing, and economic data

ACS releases include:

• One-year estimates for populations above 65,000

• Five-year estimates for smaller geographies such as census tracts and small towns

Golden emphasized that the ACS now provides:

• 20 years of historical data

allowing researchers to study long-term patterns such as persistent county-level poverty.

She referenced ACS-based work showing concentrated persistent poverty in parts of the United States, noting that Bronx County was the primary New York county appearing prominently in those analyses.

Survey of Income and Program Participation (SIPP)

Golden described SIPP as a longitudinal survey that follows the same households over time.

Key characteristics include:

• Began in 1983

• Approximately 50,000 sampled addresses

• Tracks respondents for roughly four years

SIPP is especially useful for studying:

• Chronic poverty

• Episodic poverty

• Program participation over time

Golden explained that SIPP can identify households that move in and out of poverty over time rather than remaining continuously poor.

Small Area Income and Poverty Estimates (SAIPE)

Golden then introduced SAIPE, which provides:

• Model-based poverty estimates

• Data for states, counties, and school districts

• Estimates from 1995 onward

SAIPE combines:

• ACS data

• Administrative data

to generate more granular poverty estimates.

The dataset is particularly important because:

• The U.S. Department of Education uses it to allocate Title I funding

Golden demonstrated the SAIPE tool live, filtering:

• New York State counties

• Under-18 poverty rates

• County-level poverty counts

She showed that:

• Kings County and Bronx County had especially large numbers of children living in poverty

according to SAIPE estimates.

Using data.census.gov

A substantial portion of the session focused on demonstrating data.census.gov, the Census Bureau’s primary public data portal.

Golden described the platform as:

• A central repository for Census data

• Supporting tables, maps, charts, downloads, APIs, and geospatial analysis

She walked through a live example using:

• ACS five-year estimates

• Table S1701 (poverty status in the past 12 months)

• County-level poverty data for New York State

Golden demonstrated how users can:

• Filter by geography

• Filter by dataset

• Remove margins of error

• Map poverty rates

• Generate charts

• Change variables dynamically

She highlighted Bronx County as having the highest poverty percentage among New York counties in the demonstration map at approximately 27.8%.

Mapping SNAP Data

Golden also demonstrated how users can map SNAP participation data at the census tract level using data.census.gov.

Using an address near Chinatown in Manhattan, she showed how users can:

• Select census tracts interactively

• Overlay boundaries such as community districts

• Map SNAP participation variables

• Compare neighborhood-level poverty-related indicators

Audience Questions – Census Data Infrastructure

During Q&A, participants asked about the infrastructure underlying data.census.gov.

Golden explained that the platform is a proprietary Census Bureau-built system, though it may incorporate some off-the-shelf software components.

She contrasted it with older Census tools such as:

• American FactFinder

• DataFerrett

which have now largely been replaced by data.census.gov.

Audience Questions – Poverty Table “Cheat Sheets”

Audience members also asked whether a master list or “cheat sheet” exists for all poverty tables.

Golden and Glassman explained that:

• No single exhaustive poverty-table cheat sheet exists

• Users can search tables by subject or keywords within data.census.gov

• Census staff are available to assist researchers seeking specific poverty tables

Closing Remarks

Golden concluded by encouraging organizations to contact Census Bureau staff for customized trainings, presentations, and support related to Census data and poverty statistics.

RESOURCES

• data.census.gov — the Census Bureau’s main data platform for maps, charts, and tables, demonstrated live by Joli Golden

• Census Bureau Poverty topics page — central hub for poverty estimates, surveys, and guidance on choosing the right data source

• Supplemental Poverty Measure (SPM) — the measure that factors in tax credits, government assistance, geography, and key expenses

• Small Area Income and Poverty Estimates (SAIPE) Program — model-based poverty estimates for states, counties, and school districts, used for Title I allocations

• SAIPE Interactive Data Tool — the interactive application Joli demonstrated for exploring county and school-district poverty data

• Survey of Income and Program Participation (SIPP) — the longitudinal survey used to measure chronic and episodic poverty over time

• Poverty in the United States: 2024 — the official poverty report drawn from the CPS ASEC, produced by Brian Glassman’s branch

• Many U.S. Counties Had High Poverty Rates Over 20 Years — America Counts story by Craig Benson, referenced during the ACS discussion

• Census Academy — free courses, Data Gems, and recorded webinars on using Census Bureau data

• Exploring the Microdata Access Tool (MDAT) — video tutorial on the MDAT, the advanced microdata tool Joli linked for the audience

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Jan Mendez - Budget and Policy Analyst, New York City Independent Budget Office (IBO); Brian Cain - Assistant Director of Housing, Environment, and Infrastructure, IBO; Gytis Simaitis - Fox Hall Analytics
Moderator: Brian Cain - IBO

Introduction to the Independent Budget Office and the Bus Speed Study

Jan Mendez introduced himself as a budget and policy analyst with the New York City Independent Budget Office (IBO), working within the Housing, Environment, and Infrastructure team. He explained that IBO is a nonpartisan city agency that provides data-driven analysis of how New York City raises and spends money, including evaluations of major policy issues affecting everyday life in the city.

Mendez said the office became interested in buses because public transportation is essential to the functioning of New York City, particularly for:

• Residents without subway access

• People with disabilities

• Communities heavily dependent on bus service

He noted that buses are frequently perceived as “slow and unreliable,” a sentiment widely shared by riders and commonly discussed in public conversations about transit. IBO therefore wanted to investigate whether that perception was supported by data and what underlying factors might explain it.

National Context – NYC Buses Among the Slowest in the United States

Mendez reviewed prior findings from a 2017 report by the NYC Comptroller’s Office, which concluded that New York City buses averaged approximately 7.4 miles per hour and were among the slowest bus systems in the country.

IBO conducted its own analysis using Federal Transit Administration data and found that New York City bus divisions continue to rank among the slowest major bus systems in the United States.

Comparisons with other systems showed:

• MTA bus divisions lagging behind peer cities

• New Jersey Transit operating substantially faster than NYC buses

• Local and Select Bus Service (SBS) routes remaining particularly slow

Mendez emphasized that this was not simply anecdotal frustration but a measurable structural issue within the city’s transportation system.

Key Open Data Sources Used in the Study

The presentation focused heavily on the role of publicly available NYC Open Data in enabling the analysis.

IBO relied on three primary datasets:

• MTA Bus Speeds dataset

• NYC DOT Bus Lanes dataset

• NYC Department of Finance Parking Violations dataset

Mendez highlighted that all of these datasets are publicly accessible and relatively easy to use, making them valuable tools not only for government analysts but also for researchers, advocates, and members of the public.

MTA Bus Speeds Dataset

The MTA bus speeds dataset was described as the core foundation of the analysis. The dataset is updated monthly and includes:

• Average bus speeds

• Total operating time

• Total mileage

• Route-level information

dating back to January 2015.

The data can be filtered by:

• Borough

• Weekday versus weekend

• Peak versus off-peak hours

• Local buses

• SBS buses

• Express buses

Mendez praised the dataset as “extremely easy to use” and encouraged attendees to explore it themselves.

NYC DOT Bus Lane Dataset

The second major dataset mapped all bus lanes in New York City. It included information such as:

• Location

• Direction

• Date built

• Width

• Borough

• GIS identifiers linked to the LION street dataset

Mendez noted that the dataset worked particularly well with GIS tools such as ArcGIS Pro and QGIS.

However, one important limitation emerged:

The dataset did not distinguish between protected and unprotected bus lanes.

This limitation became especially important later when evaluating compliance with the NYC Streets Plan, which specifically requires “protected” bus lanes.

Parking Violations Dataset

The Department of Finance parking violations dataset was used to analyze enforcement patterns affecting bus movement.

Mendez explained that the dataset:

• Is organized by fiscal year

• Contains extremely large files

• Requires significant cleaning and filtering

IBO used R to isolate bus-related violation codes involving:

• Bus lane blocking

• Standing in bus stops

• Failure to turn from bus lanes

• Mobile bus lane enforcement

Analytical Methodology

To minimize distortions, IBO focused primarily on:

• Weekday peak-hour trips

• Pre-pandemic baseline comparisons

• January 2019 as a benchmark

The office intentionally excluded weekend and off-peak variability because weekday commuting periods most strongly reflected riders’ lived experiences with slow buses.

Mendez explained that COVID-era conditions temporarily increased bus speeds because:

• Ridership collapsed

• Traffic volumes fell

• Overall congestion decreased

As the city reopened, speeds largely returned to earlier patterns.

Findings – Bus Speeds Remain Chronically Slow

IBO found that average bus speeds remain very close to the levels identified years earlier by the Comptroller’s Office.

Key findings included:

• Express buses remained the fastest category

• SBS buses were only modestly faster than local buses

• Manhattan buses were the slowest overall

• Staten Island buses were the fastest overall

Mendez emphasized that Manhattan’s dense street network and heavy congestion strongly contributed to slow travel times.

The data also showed recurring seasonal fluctuations:

• Summer speeds differed from winter and spring patterns

• Weekend buses were somewhat faster than weekday buses

• Express buses showed the greatest weekend speed improvements

“The Slowest Bus in NYC”

Mendez then presented several memorable route-level statistics.

For 2025:

• The slowest bus route in NYC was the Manhattan M57

• Average speed: approximately 4.99 mph

He joked that some people could jog faster than the bus.

Other notable findings included:

• Fastest non-express bus: Staten Island S89 at approximately 16.79 mph

• Fastest Manhattan non-express bus: M98 at approximately 9.19 mph

• Fastest overall bus: SIM24 express route at nearly 25 mph

One striking comparison showed that:

• The slowest Staten Island bus was still faster than the fastest Manhattan bus

Major Causes of Slow Bus Speeds

IBO identified three major contributors to slow bus performance.

Traffic Congestion

The most visible issue was severe traffic congestion, particularly in Manhattan.

Mendez cited MTA estimates that:

• More than 700,000 people enter Manhattan’s central business district daily

Heavy congestion reduces travel speeds for all vehicles, including buses.

Lack of Dedicated Bus Lanes

Dedicated bus lanes were identified as another major factor.

Mendez explained that bus lanes allow buses to continue moving even when adjacent traffic is heavily congested. Major corridors such as Fifth Avenue and 42nd Street use these lanes to improve reliability and speed.

Weak Enforcement of Bus Lane Rules

The third major issue involved inconsistent enforcement.

IBO examined parking violation data and found that:

• Camera-based bus lane violations increased sharply after 2020

• Officer-issued bus lane violations plateaued after 2023

Mendez explained that automated camera enforcement had expanded rapidly and now plays a much larger role than traditional officer-issued tickets.

The NYC Streets Plan

The presentation then shifted toward the city’s policy response.

Mendez explained that Local Law 195 of 2019 required the NYC Department of Transportation (DOT) to implement the NYC Streets Plan — a five-year transportation master plan covering:

• Bus lanes

• Bike lanes

• Pedestrian infrastructure

• Transit signal priority

• Accessibility improvements

One of the plan’s major mandates was:

• Construction of 150 miles of protected bus lanes over five years

This represented an extremely ambitious target because it effectively required DOT to double the total historical mileage of bus lanes built in New York City.

Protected vs. Unprotected Bus Lanes

Mendez clarified an important distinction between ordinary bus lanes and “protected” bus lanes.

Ordinary bus lanes:

• Consist mainly of painted roadway markings

Protected bus lanes:

• Must either be physically separated from traffic

• Or equipped with automated camera enforcement

This distinction significantly affected Streets Plan compliance calculations.

Streets Plan Progress and Shortfalls

IBO found that DOT had achieved strong progress in some Streets Plan categories, including:

• Redesigned intersections

• Pedestrian spaces

• Accessible pedestrian signals

• Transit priority signals

However, bus lane construction lagged badly.

By 2025:

• Only 44 miles of protected bus lanes had been completed

• 106 additional miles would still need to be built in 2026 to meet the 150-mile mandate

Mendez openly questioned whether that target was realistically achievable.

Geographic Distribution of Bus Lane Construction

Using GIS mapping, IBO analyzed where new bus lanes had been added.

Findings showed:

• Manhattan had the highest density of bus lanes overall

• Bronx and Manhattan received the greatest recent expansions

• Staten Island received no new bus lanes after the Streets Plan began

New investments in Manhattan focused on corridors including:

• Third Avenue

• 96th Street

• Lower East Side routes

Bronx investments focused on:

• University Avenue

• Mosholu Parkway

• East Gun Hill Road

Why DOT Is Struggling to Meet the Mandates

Mendez summarized several reasons DOT has struggled to meet Streets Plan requirements.

Funding Constraints

DOT has repeatedly argued that achieving the mandates would require:

• Billions in additional funding

• Major staffing increases

• Significantly expanded capital budgets

City Council estimates projected:

• Approximately $377 million in additional expense funding

• Approximately $252 million in additional capital funding

would be required to meet the minimum mandates.

Strict Definitions

Another challenge involved the legal definitions embedded in the Streets Plan.

Simply painting new bus lanes does not satisfy the law unless they qualify as protected lanes.

Political Resistance

Mendez also cited opposition from:

• Community boards

• Local elected officials

• Neighborhood stakeholders

who have resisted some proposed street redesigns.

Social and Economic Impacts of Slow Buses

Mendez repeatedly emphasized that bus speeds are not “just numbers.”

Slow and unreliable buses can:

• Limit access to jobs

• Increase stress and anxiety

• Reduce mobility

• Harm economic development

• Affect lower-income communities disproportionately

IBO’s analysis found that bus riders generally have:

• Lower median incomes than subway riders

• Lower incomes than car commuters

Slow buses therefore disproportionately burden already vulnerable populations.

Environmental Impacts

The presentation also highlighted environmental concerns.

Buses stuck in traffic:

• Idle longer

• Burn more fuel

• Produce additional pollution

Mendez acknowledged that the MTA has made substantial progress toward environmental goals but argued that congestion remains a major environmental issue.

Recent Developments – Congestion Pricing and New Data

Several major updates had occurred after the report’s publication.

Congestion Pricing

Congestion pricing began operating in Manhattan’s central business district.

According to MTA data:

• Bus speeds inside the congestion pricing zone increased approximately 2.3%

Segment-Level Bus Speed Dataset

The MTA also released a new segment-level speed dataset that breaks routes into smaller sections instead of reporting only route-wide averages.

This allows analysts to identify:

• Specific bottlenecks

• Localized slow segments

• Midtown-specific delays

rather than averaging conditions across entire routes.

Free Bus Proposal

Mendez also noted that newly inaugurated Mayor Zohran Mamdani campaigned on making city buses fare-free.

IBO had not yet evaluated the potential impacts, though the topic surfaced repeatedly during Q&A discussions.

Q&A – Methodology and Reliability

Brian Cain moderated a lengthy audience Q&A session.

Participants asked about:

• Average speed calculations

• Standard deviations

• Passenger boarding impacts

• Fare evasion

• SBS OMNY transitions

• Reliability versus speed

• Bus stop spacing

• Dwell time

• Segment-level GPS analysis

Mendez explained that IBO calculated average speed by dividing:

• Total operating time

• By total mileage

consistent with Federal Transit Administration methods.

He acknowledged that many additional variables — including passenger boarding volumes and stop spacing — likely influence bus speeds but were beyond the initial scope of the study.

Fare Evasion Discussion

Audience members discussed whether fare evasion contributes to delays.

Mendez noted that fare evasion rates on local buses are estimated at roughly:

• 40–50%

He acknowledged that fare disputes and enforcement interactions can delay service, though the exact impact remains difficult to quantify.

Gytis Simaitis of Fox Hall Analytics described how crowding near rear doors and fare avoidance behavior can slow boarding and movement.

Reliability Versus Speed

Audience members also raised concerns about reliability — buses failing to arrive on schedule at all.

Mendez acknowledged that reliability is a critical issue but explained that measuring it properly would require:

• Timetable comparisons

• GTFS schedule analysis

• Arrival tracking

which fell outside the scope of the initial report.

Final Conclusions

Mendez concluded that New York City buses remain persistently slow and that current policy interventions have not yet solved the underlying structural problems.

He stressed that improving bus service is essential because buses disproportionately serve:

• Older adults

• People with disabilities

• Lower-income communities

• Residents without subway access

While the Streets Plan and congestion pricing may improve conditions over time, IBO concluded that the city still faces substantial challenges in creating a reliable, efficient, and equitable bus network.

RESOURCES

• Speeding Up Slowly: A Review of Initiatives to Improve Bus Speeds in New York City — the February 2025 IBO report that this presentation is based on

• Mapping New York City’s Efforts to Improve Bus Speeds — IBO’s interactive StoryMap accompanying the report

• Bus Speeds in New York City Interactive Map — explorable map of average bus speeds by route, referenced by Jan Mendez

• NYC Independent Budget Office — Environment, Infrastructure & Capital Budget — IBO’s transportation and infrastructure publications

• Behind Schedule: How New York City’s Bus System Slow Rolls Riders — April 2025 NYC Comptroller report on slow buses and congestion pricing impacts

• NYC Streets Plan — NYC DOT’s five-year transportation master plan and annual progress updates

• Bus Lanes - Local Streets — NYC Open Data dataset of bus lanes used in IBO’s analysis

• MTA Bus Route Segment Speeds: Beginning 2025 — segment-level bus speed data discussed as a recent MTA release

• How NYC Open Data Guided a Review of Initiatives to Improve Bus Speeds in New York City — the NYC Open Data Week 2026 event page

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: David Tussey - Former Tech Executive, NYC
Moderator: Aleksandr Finkel - NYC OTI

Introduction to the 311 Quality of Life Analysis

David Tussey opened the session by explaining that the presentation examined whether New York City’s massive 311 complaint database could be used to identify broader trends related to urban quality of life. Tussey, a former technology executive for New York City government and former leader within what was then DoITT, explained that he had participated in a major modernization of the city’s 311 software platform in 2020.

He described 311 as the city’s non-emergency service and complaint system, contrasting it with 911 emergency services. Approximately 3 million 311 complaints are filed annually through:

• Phone calls

• Online submissions

• Mobile applications

• Other digital channels

Tussey framed the central research question as:

Can 311 complaints function as a proxy or “barometer” for understanding changes in New York City residents’ quality of life?

He noted that while crime statistics often dominate public discussion about urban conditions, many non-criminal issues — such as sanitation, noise, graffiti, or infrastructure maintenance — also strongly shape everyday urban experience.

Initial Research Design and Unexpected Data Challenges

Tussey originally planned a relatively simple longitudinal comparison using 2021 complaint data as a baseline and comparing subsequent years during Mayor Eric Adams’ administration.

However, the project quickly became more complicated because the 2020–2021 data proved highly unstable.

Problems included:

• Missing records

• Large unexplained spikes

• COVID-era distortions

• System transition effects from the new 311 platform

Tussey explained that many complaint categories collapsed during COVID lockdowns while others surged dramatically, making the data unsuitable for establishing a stable baseline.

As a result, he shifted the baseline year to 2022 and ultimately analyzed trends across 2023–2025 using a seasonally adjusted index system.

The seasonally adjusted approach normalized complaint volumes relative to expected complaint levels for each month. For example:

• A score of 1.5 in December meant complaints were 1.5 times higher than typical December complaint levels

• A score of 2.0 meant complaints doubled expected seasonal norms

Tussey emphasized that seasonal normalization was necessary because many complaint categories vary dramatically by weather and time of year.

Technical Workflow and Use of Open Data

Tussey described the project’s technical architecture in detail.

He downloaded approximately four years of 311 data directly from NYC Open Data as a flat CSV file containing:

• Approximately 13.5 million complaint records

• Roughly 1 gigabyte of data

The analysis pipeline was built in R, and Tussey noted that he used Anthropic’s Claude AI system extensively to accelerate software development and automate coding tasks.

Key preprocessing steps included:

• Converting text date fields into actual date formats

• Normalizing inconsistent capitalization

• Standardizing column naming conventions

• Converting the data into RDS format for more efficient processing

Tussey stressed that data cleaning consumed a substantial portion of the project effort and described dirty or inconsistent data as one of the largest challenges in any large-scale open data analysis project.

Creating “Quality of Life Indicators”

The original 311 dataset contained approximately 241 complaint categories. Tussey manually selected about 60 categories that he believed reflected quality-of-life conditions.

Example categories included:

• Graffiti

• Homeless assistance

• Standing water

• Rodent sightings

• Drug activity

• Broken parking meters

• Illegal dumping

• Unsanitary conditions

Because 60 indicators proved too unwieldy, he grouped related complaint types into “families,” reducing the number to 36 broader groupings.

Examples included:

• Animal-related complaints

• Unsanitary conditions

• Street safety issues

• Sanitation complaints

He then grouped the families into even larger “bundles,” creating 11 executive-level heat maps intended to summarize broad categories of urban quality-of-life conditions.

Tussey repeatedly emphasized that these groupings were subjective and open to revision.

Standing Water and Seasonal Complaint Cycles

Tussey used standing water complaints to demonstrate how seasonal trend analysis worked.

Standing water complaints — typically involving blocked drains or roadway flooding — showed strong seasonal patterns:

• Very high complaint volumes during spring and summer

• Near-zero complaints during winter months

Charts displayed recurring annual “sawtooth” cycles corresponding to seasonal weather changes. Tussey highlighted that understanding these recurring cycles was essential before attempting to identify abnormal trends.

Graffiti Complaints and Urban Blight

Graffiti complaints were presented as an example of a worsening quality-of-life indicator.

Tussey showed that:

• Graffiti complaints followed a recurring spring/summer increase

• Overall complaint trends were rising substantially over time

The analysis estimated graffiti complaints were:

• Up approximately 58% compared with the 2022 baseline

• Associated with nearly 60,000 complaints across the four-year period

Tussey interpreted this as evidence of worsening urban blight conditions.

Discussion on Bias and Interpretation

During the session, audience member Sherry questioned whether categories such as “homeless assistance” reflected a bias toward the concerns of housed residents rather than unhoused populations themselves.

Tussey acknowledged the criticism and agreed that the selection of quality-of-life indicators was inherently subjective. He stated that the project should be understood as exploratory and open to alternative interpretations and categorization systems.

Positive Trend Indicators

Tussey then shifted to several complaint categories showing improvement.

Consumer Complaints

Consumer complaints — including complaints against:

• Retail stores

• Bodegas

• Dry cleaners

• Towing companies

• Parking garages

showed a substantial decline after a spike in 2022.

The analysis estimated:

• Consumer complaints declined approximately 21%

over the observed period.

Streetlight Condition Complaints

Streetlight and traffic signal complaints also declined substantially:

• Down approximately 40%

Tussey interpreted this as evidence of improved infrastructure maintenance.

Rodent Complaints and the “Rat Czar”

One of the session’s most discussed examples involved rodent complaints.

Tussey showed that citizen-reported rodent sightings declined approximately 25% over the study period.

Audience members suggested several explanations:

• The city’s new sealed trash bin program

• Increased sanitation enforcement

• The appointment of the city’s “rat czar”

• Expanded rodent mitigation efforts

Tussey credited Department of Sanitation leadership under Jessica Tisch and argued that the data appeared to demonstrate measurable policy success.

Broken Parking Meters

Complaints about broken parking meters declined approximately 40%.

Participants attributed this largely to the transition away from:

• Mechanical coin-operated meters

• Toward app-based digital parking systems

Tussey cited this as an example where technological modernization directly altered complaint patterns.

Neutral or Stable Trends

Several complaint categories remained relatively stable.

Street Sweeping Complaints

Alternate-side parking and street sweeping complaints showed little overall change from 2022 levels.

Abandoned Vehicles

Abandoned vehicle complaints displayed dramatic seasonal spikes every January but no clear long-term upward or downward trend.

Participants speculated that seasonal travel patterns and “snowbird” migration might contribute to the annual January spikes.

Trash Disposal Complaints

Trash disposal complaints fluctuated heavily but ultimately showed almost no net change overall.

Tussey noted that explaining the causes of these spikes would require a much deeper second-order analysis.

Worsening Trends and Emerging Concerns

Tussey then reviewed complaint categories showing significant deterioration.

Illegal Posting and Dumping

Complaints involving illegal posting and dumping rose approximately 43%.

Tussey suggested that these trends reflected worsening neighborhood disorder conditions.

Unsanitary Conditions

A family of complaints involving:

• Bad smells

• Dirty conditions

• Trash accumulation

• General sanitation problems

increased approximately 26%.

Drug Activity Complaints

Drug activity complaints represented one of the most dramatic increases.

Tussey displayed trend points reaching:

• 8–9 times expected seasonal norms during parts of 2025

Overall:

• Drug activity complaints increased more than 300%

during the study period.

Tussey noted that these reports reflected citizen complaints rather than verified police incidents.

Audience discussion considered several possible factors:

• Cannabis legalization

• Changes in drug enforcement policy

• Shifting public attitudes toward reporting

Tussey stressed that the dataset alone could identify trends but not definitively explain causation.

Lead Complaints

Tussey identified lead-related complaints as especially troubling.

The category combined:

• Housing Preservation and Development (HPD) complaints

• Department of Buildings complaints

related to:

• Lead paint

• Lead pipes

• Lead hazards

The trend analysis suggested:

• Lead complaints increased nearly 120%

• Approximately 55,000 reports were recorded

Tussey noted that two extreme spikes exceeded the chart’s vertical scale entirely.

Audience members pointed out that expanded mandatory lead testing laws may partly explain the increase.

Executive-Level “Heat Maps”

Tussey introduced a series of executive-style heat maps designed to summarize broad issue categories visually.

Categories included:

• Public health

• Street safety

• Blight and nuisance

• Sanitation

• Transportation

Each heat map color-coded trends as:

• Improved

• Neutral

• Slightly worse

• Much worse

Examples included:

• Strong improvements in rodent complaints and parking meters

• Significant deterioration in drug activity and illegal dumping

Tussey suggested that such visualizations could potentially support:

• Budget prioritization

• Policy development

• Executive decision-making

• Community planning

Overall Findings

Tussey summarized the full set of quality-of-life indicators:

• Approximately 32% showed improvement or stability

• Approximately 27% showed slight deterioration

• Approximately 41% showed substantial worsening

He concluded that a large share of the analyzed quality-of-life indicators appeared to be deteriorating and “needed attention.”

Audience Discussion on Data Interpretation and Policy Use

The discussion section became one of the session’s most substantive components.

Audience members raised questions involving:

• Statistical methodology

• Mean versus median normalization

• Seasonal adjustment

• Repeat complainants

• “Complaint fatigue”

• Free-text parsing

• Geographic mapping

Tussey acknowledged many methodological limitations, including:

• Reliance on a single baseline year

• Subjective category selection

• Incomplete causal analysis

He emphasized that the project should be viewed as exploratory rather than definitive.

Geographic and Agency-Level Analysis

Tussey explained that the underlying 311 data include:

• Street addresses

• Latitude/longitude coordinates

• Responsible city agencies

This makes it possible to perform:

• Borough-level comparisons

• Community board analysis

• Agency performance comparisons

• Geospatial mapping

He briefly displayed borough-level complaint trends showing:

• Manhattan complaints up roughly 24%

• Queens up roughly 17%

• Staten Island mostly unchanged

• Bronx slightly down

Tussey cautioned that complaint counts become statistically sparse at smaller geographic scales.

Discussion on Tenant Harassment and Hidden Patterns

One of the most important audience exchanges involved Virginia Crawford, who discussed difficulties tracking landlord harassment against rent-stabilized tenants.

Crawford explained that:

• 311 does not explicitly classify harassment complaints

• Harassment cases must often be reconstructed from housing court data

• Many individual complaints (rodents, broken locks, mail theft, construction issues) may collectively indicate systematic tenant harassment

She proposed grouping seemingly unrelated complaints into “families” representing broader structural housing harassment patterns.

Tussey strongly supported the idea and argued that this was precisely the type of higher-order policy analysis the city should pursue using integrated datasets.

The Role of 311 as Civic Infrastructure

Sherry emphasized that residents are constantly encouraged by community boards and city officials to submit 311 complaints as a way of documenting neighborhood problems and attracting government attention.

Tussey agreed and argued that because residents increasingly use 311 as a civic reporting mechanism, the city has a responsibility not only to collect the data but also to analyze and operationalize it effectively.

He stressed that without meaningful analysis:

“We’re just collecting data.”

Historical Origins of NYC 311

Toward the end of the session, Tussey discussed the origins of NYC 311 under Mayor Michael Bloomberg.

He explained that Bloomberg’s administration consolidated numerous fragmented agency hotlines into a centralized service system inspired partly by Bloomberg LP’s centralized global support model.

Tussey also noted that the major 2019–2020 software modernization was the first large-scale platform replacement since the original 2003 implementation.

Final Reflections on Open Data and Civic Analysis

Tussey concluded by arguing that the true value of open data lies not simply in publication, but in interpretation and civic use.

He advocated for:

• Dedicated city analytical teams

• Better integration of datasets

• Executive dashboards

• Community-level report cards

• More active use of data in policy formation

Tussey repeatedly emphasized that the 311 system contains an extraordinarily rich civic dataset capable of revealing:

• Infrastructure failures

• Social stress

• Neighborhood decline

• Policy success

• Emerging urban issues

if properly analyzed and integrated into decision-making processes.

RESOURCES

• What 13 Million 311 Complaints Reveal About NYC’s Quality of Life — the NYC Open Data Week 2026 event page, with full description and event materials

• Principles for Open Data Curation: A Case Study with the NYC 311 Service Request Data — the peer-reviewed paper by David Tussey and Jun Yan underpinning this analysis

• 311 Service Requests from 2020 to Present — the NYC Open Data dataset of ~13 million complaint records analyzed in the talk

• NYC Open Data — the City’s open data portal where the 311 four-year export was obtained

• NYC Office of Technology and Innovation — Open Data — OTI (formerly DoITT) runs 311 and the Open Data program

• University of Connecticut Department of Statistics — academic home of Dr. Jun Yan, who mentored the statistical methodology

• The R Project for Statistical Computing — the language used to build the data pipeline, indexing, and trend charts

• DSNY Waste Containerization — the trash bin program credited in the session for the drop in rodent complaints

• BetaNYC — civic-tech organization referenced for neighborhood report cards and open data advocacy

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Lauren Drumgold - Policy Advisor, NYC Mayor’s Office of Food Policy; Rositsa T. Ilieva - Director of Policy, CUNY Urban Food Policy Institute; Yvette Ng - Research Fellow, CUNY Urban Food Policy Institute
Moderator: Michael Xie - NYC Open Data, NYC OTI

Introduction to the Mayor’s Office of Food Policy and Food Forward

Lauren Drumgold opened the session by introducing the NYC Mayor’s Office of Food Policy and explaining its mission to help make New York City’s food system healthier, more equitable, and more sustainable.

She explained that the office’s work is guided by Food Forward, New York City’s 10-year food policy plan focused on:

• Food economy development

• Food retail and accessibility

• Food production and distribution

• Sustainable food disposal systems

Rositsa T. Ilieva then introduced the CUNY Urban Food Policy Institute, describing it as a research and action center based at the CUNY Graduate School of Public Health and Health Policy in Harlem. She emphasized the institute’s partnerships with community organizations, food advocates, and government agencies to design evidence-based urban food policy interventions.

Yvette Ng briefly introduced herself as a research fellow at the institute who had previously worked on dashboard development projects and was now part of the team building the new interactive food policy dashboard.

Evolution of NYC Food Policy Reporting

Drumgold reviewed the evolution of the city’s food policy reporting systems.

From 2012 through 2022, the city primarily relied on the annual Food Metrics Report, a narrative-style publication compiling food system data from city agencies and broader food-related initiatives.

The report covered topics including:

• Food insecurity

• SNAP participation

• Emergency food distribution

• Food procurement

• Food standards compliance

• Nutrition programs

• Farmer’s markets

• Green Carts

• Food retail access

• Urban agriculture

• Food manufacturing

• Food waste programs

Drumgold explained that although the reports contained extensive information, the presentation format had major limitations:

• Static charts and maps

• Large spreadsheet appendices

• Difficult-to-analyze tables

• Limited usability for researchers and advocates

She demonstrated how important metrics — such as funding levels, acreage, or participating farms — were embedded within text-heavy spreadsheet tables that required substantial manual cleaning before meaningful analysis could occur.

Feedback from advocates and academics pushed the office to rethink its reporting structure and improve public accessibility and data transparency.

Transition to “Food by the Numbers”

In 2023, the office launched Food by the Numbers, a more visually oriented infographic-style report designed to present major food policy metrics more clearly and concisely.

The newer format emphasized:

• Visual storytelling

• Quick statistics

• Simplified graphics

• Easier public comprehension

Examples shown during the presentation included:

• School food education grant metrics

• Nutrition incentive program participation

• Grocery access initiatives

• FRESH supermarket development

• Shop Healthy NYC programs

Drumgold explained that while the newer reports improved visual accessibility, they still remained fundamentally static documents released annually.

Vision for the Interactive Food Policy Dashboard

The core focus of the session was the city’s upcoming interactive food policy dashboard, being developed collaboratively by the Mayor’s Office of Food Policy and the CUNY Urban Food Policy Institute.

Drumgold described the dashboard as a major modernization effort intended to move city food policy reporting:

“from static to dynamic.”

The dashboard was anticipated for public release in summer or early fall 2026.

Key goals included:

• Improving public understanding of NYC’s food system

• Increasing data transparency

• Providing downloadable datasets

• Supporting policy analysis

• Assisting researchers and advocates

• Enabling interactive exploration of trends over time

The project relies heavily on publicly available datasets, especially those hosted on NYC Open Data.

Dashboard Architecture and Core Themes

Rositsa Ilieva explained that the dashboard homepage would function as a centralized “one-stop shop” for exploring the city’s food system.

The structure aligns with major themes from prior Food Metrics Reports while introducing dynamic interactive functionality.

Core sections include:

• Food insecurity and food assistance

• Meals served through city agencies

• Healthy food access and retail environment

• Food affordability

• Food and climate

• Nutrition and health outcomes

Ilieva emphasized that the dashboard integrates datasets scattered across multiple agencies into a unified interface where users can:

• Filter data

• Zoom into maps

• Compare geographies

• Explore trends over time

• Download underlying datasets

Food Insecurity and Food Assistance Section

Ilieva previewed the food insecurity and food assistance section, which integrates data from:

• SNAP

• WIC

• Community Food Connection

• Food insecurity estimates

One preview visualization displayed:

• Trends in pantry visits

• Soup kitchen meal service

• Counts of emergency food providers

• Geographic distribution of providers

The dashboard combines:

• Time-series charts

• Geographic maps

• Key statistics

• Program metrics

allowing users to examine both scale and spatial distribution simultaneously.

Meals Served Through City Agencies

The meals section highlights the approximately 220 million meals and snacks served annually across 11 city agencies.

The dashboard includes agency-specific visualizations for organizations such as:

• NYC Department of Education

• Department for the Aging

and allows users to explore:

• School breakfast participation

• Lunch participation

• After-school meals

• Borough-level meal distribution

• Longitudinal trends

Ilieva stressed that the dashboard aims to connect citywide totals with more granular operational data.

Healthy Food Access and Retail Environment

The dashboard’s food access section combines information about:

• Farmer’s markets

• Food retail incentives

• Nutrition incentive programs

• Grocery access initiatives

One featured example showed:

• Farmer’s market locations

• SNAP household concentrations by community district

• Year-round versus seasonal markets

• EBT acceptance

• Health Bucks participation

Ilieva emphasized that layering neighborhood demographic information alongside food access data helps create a more equity-focused understanding of food retail environments.

Food Affordability Metrics

The affordability section focuses on the growing gap between food prices and household purchasing power.

Visualizations include:

• Meal cost estimates

• Borough-level affordability differences

• Trends in food cost growth

• Estimates of additional income needed by food insecure households

Ilieva framed affordability not simply as a pricing issue, but as a broader equity challenge tied to economic conditions across the city.

Food and Climate Section

The climate section links food systems to sustainability and greenhouse gas emissions.

The dashboard tracks:

• Organic waste diversion

• Food waste collection

• Plant-based meal programs

• Sustainability indicators

Users will be able to filter food waste data geographically and examine trends over time using NYC Open Data datasets.

Nutrition and Health Outcomes

The health section connects food environments with health disparities and diet-related conditions.

Preview maps displayed borough-level variation in:

• Fruit and vegetable consumption

• Sugary drink intake

Ilieva explained that these visualizations allow users to identify disparities and investigate relationships between food access and health outcomes.

Live Demo – SNAP Dashboard Functionality

Yvette Ng then conducted a live walkthrough of the dashboard’s food insecurity and assistance section.

The SNAP page uses NYC Open Data from the Department of Social Services combined with 2020 Census data.

Users can dynamically switch between metrics including:

• Percent of population receiving SNAP

• Number of households

• Number of participants

Interactive maps color-code community districts according to SNAP participation rates, revealing particularly high participation levels in parts of the Bronx and Brooklyn.

Ng showed how users can:

• Filter by borough

• Drill down into community districts

• Explore longitudinal participation trends

• Compare local and citywide patterns

One example highlighted Brooklyn Community District 304, which showed declining SNAP participation over time — potentially indicating either reduced need or unmet outreach challenges.

SNAP Outreach Metrics

The dashboard also includes information on SNAP outreach efforts and the USDA Program Access Index (PAI), which measures SNAP participation among income-eligible populations.

The city historically performs above the national average on the PAI metric, though Ng noted that the Department of Social Services estimates that approximately 20% of eligible residents remain unenrolled.

The dashboard tracks:

• Outreach events

• Community partnerships

• Client engagement metrics

WIC Dashboard Section

Ng demonstrated the WIC section, which provides:

• Participant totals

• Participant breakdowns by women, infants, and children

• Borough comparisons

• WIC agency locations

• WIC retailer maps

She explained that future iterations may incorporate contextual demographic layers such as poverty rates and food insecurity indicators to better identify underserved neighborhoods.

Community Food Connection Visualizations

The Community Food Connection section visualizes:

• Food pantry visits

• Soup kitchen meals served

• Pantry and soup kitchen locations

• Borough-level service counts

Users can interactively filter maps by:

• Borough

• Facility type

• Specific provider locations

Ng emphasized that adding contextual background layers would help policymakers and advocates better evaluate whether food providers align with neighborhood need.

Food Insecurity Trends and Racial Disparities

The dashboard incorporates Feeding America’s Map the Meal Gap estimates to track food insecurity rates over time.

Key findings shown included:

• Approximately 17% citywide food insecurity

• Roughly 1.4 million food insecure residents

• Approximately 422,000 food insecure children

Ng highlighted several broader trends:

• Food insecurity has increased since 2021

• Children experience higher food insecurity rates than the overall population

• Racial and ethnic disparities are widening over time

The dashboard enables community organizations to use these data for:

• Needs assessments

• Advocacy

• Grant writing

• Policy analysis

Mentimeter Feedback Session

The presenters then launched an interactive Mentimeter session to gather public feedback.

Participants identified themselves primarily as representatives from:

• Nonprofits

• Government agencies

• Research and education sectors

The audience strongly agreed that they would use a dashboard with this level of detail and identified likely use cases including:

• Research

• Policy analysis

• Advocacy

• Community planning

• Grant writing

Audience Requests for Additional Data

Participants requested additional data layers and features, including:

• Grocery pricing information

• Food quality metrics

• Local food sourcing

• Diet-related illness data

• Disability-disaggregated data

• Age-disaggregated data

• Better data source documentation

• Downloadable datasets

• Storytelling features

Drumgold confirmed that the team intends to include:

• Downloadable underlying datasets

• Source citations

• Last-updated timestamps

• Narrative “data stories”

to contextualize visualizations and provide richer public engagement.

Development Timeline and Prior Work

Responding to a question from Rabia, Ilieva explained that the dashboard work builds on earlier NSF-funded pilot projects beginning in 2021–2022.

The current collaboration with the Mayor’s Office intensified during the previous year as the teams expanded the dashboard beyond earlier pilot food equity metrics into a comprehensive citywide food policy platform.

Praise from NYC Department of Health

Mahana Barbadillo from the NYC Department of Health and Mental Hygiene described the session as her favorite of Open Data Week and praised the dashboard as exactly the kind of tool needed for:

• Community planning

• Policy research

• Public health advocacy

She specifically highlighted the future importance of food insecurity tracking in light of potential federal SNAP policy changes under HR 1.

Drumgold welcomed future collaboration across city agencies working to modernize public data reporting systems.

Closing Remarks

The session concluded with invitations for continued public feedback and collaboration before the dashboard’s anticipated public release later in 2026.

RESOURCES

• NYC Mayor’s Office of Food Policy — co-presenter of the dashboard; Lauren Drumgold’s office

• CUNY Urban Food Policy Institute — leading dashboard development and design; home of Rositsa Ilieva and Yvette Ng

• NYC Open Data — primary public data source powering the dashboard

• Food Forward NYC — the city’s 10-year food policy plan that the dashboard’s themes are aligned with

• SNAP & WIC benefits — federal nutrition programs featured in the dashboard’s food assistance section

• Community Food Connection — the city’s emergency food program (formerly EFAP) supporting 700+ pantries and soup kitchens

• Feeding America – Map the Meal Gap — source of the dashboard’s food insecurity estimates

• Health Bucks — $2 NYC farmers market coupons tracked in the healthy food access section

• FRESH program — Department of City Planning zoning and tax incentive for grocery store development in underserved areas

• Good Food Purchasing dashboard — MOFP’s existing city food procurement dashboard mentioned during audience Q&A

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speaker: Donnise Hurley - Senior Geographic Analyst, Population Division, NYC Department of City Planning
Moderator: Oliver Bjornsson - NYC OTI

Introduction to The Newest New Yorkers Project

Donnise Hurley introduced herself as a Senior Geographic Analyst in the NYC Department of City Planning’s Population Division and presented a preview of Chapter 5 of the forthcoming sixth edition of The Newest New Yorkers, the division’s flagship publication on New York City’s foreign-born population.

Hurley explained that the publication series dates back to 1992 and examines the city’s immigrant population through demographic, legal, and socioeconomic lenses. The upcoming 2026 edition introduces a new interactive HTML format that she personally engineered to allow readers to engage dynamically with charts, tables, and immigration data.

The new edition will later be published online through the Population Division website and in print through the City Bookstore.

Hurley framed the presentation around a key question:

How do lawful immigration pathways shape New York City’s demographic future?

She emphasized that immigration law is not merely administrative policy but a structural system that shapes:

• Population size

• Languages spoken

• Economic composition

• Family formation

• Labor force development

• Long-term demographic trends

Immigration Pathways as Demographic Infrastructure

Hurley argued that immigration law acts as a “filter” or “prism” that refracts immigrant flows into distinct legal categories. Whether someone arrives through family sponsorship, employment, diversity visas, or humanitarian programs fundamentally shapes the city’s future social and economic landscape.

She explained that understanding admission pathways helps policymakers and planners anticipate:

• Future language needs

• Service delivery requirements

• Labor market shifts

• Emerging immigrant communities

• Changes in neighborhood composition

Hurley also briefly referenced another upcoming NYC OTI event focused on American Sign Language and disability-related demographic data, emphasizing the broader importance of making all populations visible within public data systems.

Distinguishing DHS Administrative Data from ACS Data

A major portion of the presentation clarified the distinction between two key data sources:

1. Department of Homeland Security (DHS) administrative data on lawful permanent residents (LPRs)

2. American Community Survey (ACS) residential survey data

Hurley stressed that these datasets answer different questions.

DHS administrative data records the moment an immigrant is granted lawful permanent resident status (“green card” status). The data capture:

• Country of birth

• Intended destination

• Class of admission

• Timing of legal adjustment

She compared LPR data to a library invoice showing books delivered to a branch. The invoice records where books were originally shipped, but not whether they later moved elsewhere.

Similarly, LPR data records intended place of residence at the time of immigration processing, but not subsequent movement or settlement patterns.

The ACS, by contrast, functions more like an annual inventory of books currently on the shelves. It measures who actually resides in New York City at the time of the survey.

The ACS surveys approximately 3.5 million households annually and provides detailed demographic information including:

• Languages spoken

• Housing conditions

• Income

• Educational attainment

• Country of birth

• Household structure

Hurley emphasized that because ACS data are survey-based, they include margins of error and sampling uncertainty, unlike DHS administrative records.

Limitations and Delays in Immigration Data

Hurley discussed several limitations affecting immigration datasets.

Administrative processing delays can distort annual immigration totals, especially for refugees and asylum seekers whose status adjustments may occur years after arrival.

County-level LPR data for New York City are especially limited because:

• They are released more slowly than national data

• They are not retroactively revised

• Processing backlogs are not always incorporated

As a result, Hurley explained that the Population Division often aggregates immigration data across decades rather than relying heavily on single-year fluctuations.

Immigration Law and the 1965 Immigration and Nationality Act

Hurley reviewed the historical foundations of modern U.S. immigration law.

The Immigration and Nationality Act of 1965 replaced earlier quota systems that heavily favored Northern and Western European immigrants.

She described the earlier quota system as discriminatory and noted that the 1965 law aligned more closely with the broader civil rights reforms of the era.

The 1965 framework prioritized:

• Family reunification

• Employment-based immigration

• Refugee and asylum admissions

The Immigration Act of 1990 further expanded:

• Employment-based immigration

• Diversity visa programs

creating the five major pathways that continue to define U.S. immigration today:

• Family preference visas

• Immediate relative visas

• Employment visas

• Diversity visas

• Humanitarian admissions

Family-Based Immigration as the Core of NYC Immigration

Hurley explained that family reunification remains the cornerstone of New York City immigration.

Family preference visas include:

• Adult children of U.S. citizens

• Siblings of citizens

• Spouses and children of green card holders

Immediate relative visas include:

• Spouses of U.S. citizens

• Minor children of citizens

• Parents of citizens

These visas are especially important because they are not numerically capped.

Hurley showed that family-based immigration has historically dominated New York City’s immigration profile far more strongly than the national average.

In the 1980s:

• 82% of NYC immigrants entered through family-related pathways

By the 2010s:

• Family-related pathways still accounted for 77% of admissions

This remained significantly above the national average of approximately 65%.

Employment-Based Immigration

Employment visas were expanded substantially under the 1990 Immigration Act, increasing annual employment-based admissions from 54,000 to 140,000 nationally.

These visas prioritize:

• Highly skilled workers

• Advanced degree holders

• Individuals with extraordinary abilities

• Critical labor shortages

However, Hurley noted that New York City receives proportionally fewer employment-based immigrants than the nation overall.

In the 2010s:

• Only 8% of NYC immigrants entered through employment visas

• Compared to 15% nationally

Certain immigrant groups relied heavily on employment pathways, including:

• Koreans

• Filipinos

• Indians

• Taiwanese immigrants

Diversity Visa Program

Hurley devoted substantial attention to the Diversity Visa Program.

She explained that the program originated partly as a response to unintended consequences of the 1965 reforms, particularly declines in European immigration.

The modern Diversity Visa lottery provides approximately 50,000 visas annually to immigrants from countries with relatively low recent migration rates to the United States.

The system includes:

• Country caps

• Geographic balancing

• Lottery-based selection

Hurley described diversity visas as an especially important mechanism for establishing entirely new immigrant communities before large family sponsorship networks exist.

Bangladesh provided a major example.

In the 1980s:

• Nearly half of Bangladeshi immigrants entered through diversity visas

Once the Bangladeshi population became established, family-based immigration gradually replaced diversity visas as the dominant pathway.

Hurley described diversity visas as a “tool for foresight,” allowing analysts to identify emerging immigrant communities before they appear prominently in Census data.

Refugee and Humanitarian Admissions

Hurley reviewed the Refugee Act of 1980 and explained the distinction between:

• Refugees processed abroad

• Asylees applying within the U.S. or at ports of entry

Refugee admissions are capped annually by presidential determination.

Hurley noted that the 2026 refugee cap of 7,500 represented the lowest refugee ceiling in modern U.S. history.

She emphasized that refugee policy is shaped not only by humanitarian concerns but also by broader geopolitical priorities.

The presentation showed how changes in humanitarian admissions strongly affected immigrant flows from:

• China

• Ukraine

Chinese refugee admissions, for example, declined sharply between the 2000s and 2010s, contributing substantially to an overall decline in refugee flows into New York City.

Changing Geography of Immigration

Hurley demonstrated how immigration pathways have reshaped New York City’s immigrant population over the past four decades.

Key trends included:

• Continued dominance of Dominican immigration through family pathways

• Growth of Bangladeshi immigration through diversity and later family sponsorship

• Expansion of immigration from Uzbekistan and Ghana

• Declines in Trinidadian and Guyanese immigration

• Strong employment-based immigration among Koreans and Filipinos

She emphasized that immigration patterns are constantly evolving as one generation establishes the legal and social infrastructure that enables future migration.

“Pipeline of the Future” Countries

Hurley identified several countries likely to shape future immigration patterns.

Yemen experienced one of the largest recent increases in immigration flows, with admissions rising approximately 66%.

Ghana also emerged as a rapidly growing source country, rising from the 52nd largest source country in the 1980s to the 16th largest by the 2010s.

Hurley suggested that current immigration data provide an early signal of future demographic transformations across the city.

Long-Term Immigration Trends and Future Uncertainty

Hurley summarized several major conclusions:

• Nearly 1 million new permanent residents arrived in NYC during the 2010s

• NYC accounted for roughly 10% of all U.S. lawful permanent resident admissions

• Family reunification remains the dominant immigration mechanism

• Employment and humanitarian pathways remain smaller than national averages

She stressed that immigration law itself fundamentally shapes which populations enter the city and therefore shapes New York City’s demographic future.

Hurley also warned that the dramatic reduction in refugee admissions under the 2026 cap could significantly alter future flows from countries that historically relied on humanitarian pathways, especially China.

Q&A – Data Delays and Reliability

During the Q&A session, audience members asked about data timeliness and reliability.

Hurley explained that DHS collects immigration data continuously, but public release delays have increased in recent years.

For county-level NYC data:

• Delays can reach two to three years

• The most recent available data were from 2023

She emphasized that there is no evidence suggesting DHS administrative data are inaccurate, though delays and suppression thresholds create limitations.

For ACS data, Hurley reiterated that margins of error exist because the survey is sample-based rather than a full census.

Interactive Future of The Newest New Yorkers

Toward the end of the session, Hurley previewed the publication’s upcoming interactive features.

Users will be able to:

• Hover over charts for detailed values

• Sort tables dynamically

• Explore hidden totals

• Interact directly with immigration datasets

She concluded by emphasizing that there is no single immigrant experience in New York City. Different communities navigate distinct legal pathways shaped by evolving immigration laws and geopolitical conditions.

Understanding those pathways, she argued, provides critical insight into which regions and communities will shape the city’s next demographic era.

RESOURCES

• The Newest New Yorkers — flagship Population Division publication on NYC’s foreign-born population, the subject of this session’s Chapter 5/6 sneak peek

• NYC Department of City Planning — agency whose Population Division produces The Newest New Yorkers; Donnise Hurley is a Senior Geographic Analyst there

• NYC DCP Population Division — source for the demographic data and forthcoming interactive HTML edition discussed

• DHS Office of Homeland Security Statistics — LPR data — the lawful permanent resident data by state, county, and country underpinning the analysis

• American Community Survey — Census Bureau survey used as the primary source for most chapters of The Newest New Yorkers

• Immigration and Nationality Act of 1965 — landmark law that replaced the old national-origins quota system and remains the basis of current immigration policy

• Immigration Act of 1990 — expanded employment-based visas and established the permanent Diversity Visa category

• Refugee Act of 1980 — defines refugee status and the framework for refugee and asylee admissions

• Diversity Visa Program — the green card lottery discussed as a key pathway for Bangladesh, Uzbekistan, and Egypt

• NYC Mayor’s Office for People with Disabilities — partner for the upcoming OTI Beta Bagel on American Sign Language and disability data

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Young Kwon - Deputy Director of Interagency Partnerships and Language Access, NYC Mayor’s Office of Immigrant Affairs (MOIA); Erica Maurer - Assistant Director, Population Division, NYC Department of City Planning; Charles Christonikos - Geographic Analyst, Population Division, NYC Department of City Planning; Alice Castillejo - CLEAR Global
Moderator: Young Kwon - NYC Mayor’s Office of Immigrant Affairs (MOIA)

Introduction to Language Access and New York City’s Linguistic Landscape

Young Kwon opened the session by introducing the NYC Mayor’s Office of Immigrant Affairs (MOIA) language access team and framing the discussion as a continuation of the previous year’s Open Data Week session on New York City’s linguistic diversity. This year’s presentation focused less on awareness and more on practical implementation — how organizations can build, use, and own language data responsibly.

Kwon emphasized the extraordinary linguistic diversity of New York City. Drawing on work from the Endangered Language Alliance and Census data, he noted that:

• More than 700 languages are spoken in New York City

• Nearly half of New Yorkers speak a language other than English at home

• Approximately 22% of New Yorkers have limited English proficiency (LEP)

He reframed the 22% LEP statistic as approximately 1.8 million residents with active language barriers, pointing out that this population alone exceeds the total population of several major U.S. cities including Phoenix and San Antonio.

Kwon argued that language access should therefore be understood not as a niche service but as a central component of city governance and public communication.

Local Law 30 and NYC Language Access Policy

Kwon reviewed New York City’s legal framework for language access, focusing on Local Law 30 of 2017. The law requires city agencies to provide language access services and established 10 designated citywide languages using demographic data from the Census and public school system.

MOIA’s role includes:

• Monitoring language access implementation across 46 city agencies

• Providing technical assistance

• Requiring agencies to appoint language access coordinators

• Reviewing agency language access plans

• Operating a language access complaint system

• Reporting progress publicly to City Council and residents

Kwon stressed that language access should not be approached as a simple compliance checklist. Instead, MOIA frames it as a “continuous cycle of improvement” aimed at moving from compliance toward broader language justice.

The “Vicious Cycle” of Language Inequities

A central conceptual framework of the presentation involved what Kwon described as the “vicious cycle of language inequities.”

He explained that many organizations make language decisions based on assumptions rather than evidence, often defaulting to:

• Only the officially designated languages

• Existing assumptions about communities

• Lack of complaint data as proof that services are sufficient

These assumptions lead organizations to under-collect language data, which then causes language needs to remain invisible in planning and budgeting. As a result, intended populations are not reached effectively, widening equity gaps and making language needs even harder to identify.

Kwon emphasized that these failures become especially dangerous during emergencies and crises when access to timely multilingual information can directly affect health and safety outcomes.

Temporary Languages and Data-Informed Language Designation

Kwon also discussed Local Law 13 of 2023, which introduced temporary language designations. Under this framework, Wolof and Pular were designated as temporary citywide languages based on evolving demographic and migration trends.

He stressed that language selection should be intentional and data-driven rather than static or assumption-based.

NYC Department of City Planning – Population Data and Language Tools

Erica Maurer from the NYC Department of City Planning (DCP) then introduced the city’s language-related demographic tools and datasets.

Maurer reiterated several key demographic findings:

• Nearly half of New Yorkers over age five speak a language other than English at home

• Approximately 22% speak English “less than very well”

• Spanish speakers comprise nearly half of the city’s LEP population

• Chinese is the second-largest LEP language group at 19%

• Russian, Bengali, Yiddish, and Haitian Creole follow

However, Maurer emphasized that citywide averages obscure strong neighborhood variation. Different neighborhoods have very different linguistic compositions, making localized data essential for effective language access planning.

Population FactFinder and Population Map Viewer

Maurer described two major DCP tools:

• Population FactFinder (PFF)

• Population Map Viewer

Population FactFinder allows users to select geographies and explore Census and American Community Survey (ACS) data through neighborhood-level profiles. Language data are available within the ACS “Language Spoken at Home” section, including breakdowns of LEP populations by language.

Maurer explained that ACS data are survey-based and therefore include margins of error. To help users interpret the data responsibly, DCP’s tools incorporate visual “guardrails” indicating statistical reliability and significance.

The Population Map Viewer provides geographic visualizations of demographic variables across neighborhoods and community districts. Maurer noted that DCP developed a “Map Reliability Calculator” to ensure mapped ACS data are displayed responsibly and do not misrepresent uncertain estimates.

ACS Summary File vs. PUMS Data

Maurer explained the distinction between two Census data products:

1. ACS Summary File

2. Public Use Microdata Sample (PUMS)

The summary file contains larger sample sizes and allows tract-level geography, while PUMS offers much more detailed language information but only at larger geographic scales such as community districts.

DCP’s newer language datasets and tools rely heavily on PUMS because of its richer language detail.

Charles Christonikos – NYC Language Explorer Dashboard

Charles Christonikos introduced the NYC Language Explorer, a new DCP dashboard still under development and expected to launch later in 2026.

The tool is designed as two applications in one:

• Explore by Area

• Explore by Language

The “Explore by Area” mode allows users to select a geographic area and examine the linguistic composition and LEP characteristics of that area.

Features include:

• Interactive maps

• Borough and community district comparisons

• LEP percentages

• Top languages spoken

• English proficiency breakdowns

• Interactive rank-ordered charts

• Margins of error displayed directly alongside estimates

Christonikos emphasized transparency regarding uncertainty, repeatedly noting that ACS figures are estimates rather than exact counts.

Explore by Language Mode

The second mode, “Explore by Language,” focuses on individual languages across New York City.

Users can:

• Select a language such as Spanish or Chinese

• View choropleth maps showing concentrations of LEP speakers

• Examine borough-level breakdowns

• Compare language prevalence across community districts

• Analyze English proficiency within specific language groups

Christonikos explained that the tool will eventually support detailed exploration of the top 20 LEP languages in New York City, not only the official citywide languages.

He invited feedback from users regarding desired features and improvements while the application remains under development.

CLEAR Global – Digital Language Inequities and Humanitarian Communication

Alice Castillejo from CLEAR Global shifted the discussion toward international humanitarian communication and digital language inequities.

She explained that CLEAR Global works to ensure that people can access critical information and participate in decision-making regardless of the language they speak.

Castillejo described how digital communication systems and AI models depend heavily on language data availability. However, language data are distributed highly unevenly across the world.

She highlighted major disparities:

• Languages spoken in wealthier regions generally have much stronger digital representation

• Large populations such as Saraiki and Swahili speakers remain digitally underserved

• Even Arabic has significantly less digital language infrastructure than some European languages

These inequalities directly affect communities’ ability to:

• Access digital services

• Use machine translation tools

• Receive emergency information

• Participate in online systems

• Interact with AI technologies

Humanitarian Risks from Language Exclusion

Castillejo presented findings from CLEAR Global’s 2024 research showing that humanitarian digital systems are overwhelmingly:

• Text-based

• Built around dominant or official languages

This creates significant barriers for vulnerable populations, especially:

• Women

• Older adults

• People with disabilities

• Individuals with low literacy

• Speakers of marginalized languages

She described examples where people unable to understand registration systems for humanitarian aid shared personal identifying information with intermediaries, exposing themselves to exploitation and fraud.

The inability to receive emergency warnings or health information in accessible formats further increases vulnerability during disasters and epidemics.

Voice Technology and Speech Data

Castillejo argued that humanitarian communication systems increasingly need voice-based technologies rather than relying exclusively on text.

She distinguished between two kinds of data needed for speech technologies:

1. Data used to determine which languages should receive investment

2. Actual speech recordings used to build speech-recognition and text-to-speech systems

CLEAR Global combines multiple datasets into a vulnerability index that considers:

• Literacy levels

• Human development indicators

• Crisis frequency

• Humanitarian severity

• Population size

• Existing language technology quality

This system helps prioritize which languages require investment for humanitarian response systems.

Building Inclusive Speech Models

Castillejo explained that many commercial speech models fail because they are trained on the wrong dialects or unrepresentative populations.

Examples included:

• Portuguese systems trained mainly on Brazilian and European Portuguese rather than African variants

• Swahili systems trained primarily on Kenyan and Tanzanian variants rather than Congolese Swahili

She also noted that crowdsourced speech datasets often overrepresent:

• Young men

• Urban speakers

leading to weaker recognition of women’s voices, older speakers, and marginalized dialects.

CLEAR Global therefore developed a platform where volunteer linguists contribute:

• Voice recordings

• Transcriptions

• Dialect-specific data

• Domain-specific terminology

The system also allows:

• Geographic targeting

• Age balancing

• Gender balancing

• Specialized domain vocabularies

for topics such as maternal health, floods, or agriculture.

Ethical Data Collection and Community Trust

Castillejo emphasized that informed consent and ethical data stewardship are central to CLEAR Global’s model. Contributors can withdraw their data, and systems comply with European data management standards.

She also stressed that community participation itself improves trust and adoption of resulting technologies.

Final Recommendations and “Virtuous Cycles”

Returning to broader themes, Young Kwon summarized several practical recommendations:

• Start with existing datasets

• Identify data gaps critically

• Design ethically

• Prioritize strategically

• Monitor continuously

• Treat language access as iterative rather than static

Kwon argued that responsible use of language data can transform the “vicious cycle of language inequity” into a “virtuous cycle” where:

• Better data collection improves decision-making

• Resources reach intended populations

• Equity gaps narrow over time

Audience Questions on Voice Technology and AI

During Q&A, participants asked CLEAR Global for clarification on “voice data” and “voice technology.” Castillejo explained that speech technologies allow users to speak questions aloud and receive spoken responses through systems built using speech-to-text and text-to-speech models.

She emphasized that AI systems are built using human-contributed voice data rather than AI generating the data itself. Humans record and transcribe speech, which then becomes training data for machine-learning systems.

Discussion on Municipal Language Data Efforts

Audience members also asked whether other municipalities provide language data at a similar level of detail. Erica Maurer explained that ACS and PUMS data are publicly available nationwide through the Census Bureau, though relatively few cities currently provide specialized language dashboards comparable to NYC’s planned Language Explorer.

Closing Remarks

The session concluded with encouragement for attendees to continue experimenting with language data, provide feedback on the upcoming NYC Language Explorer, and think critically about how language data can be integrated ethically into service delivery, digital systems, and community engagement.

RESOURCES

• Build It, Use It, Own It — NYC Open Data Week 2026 event page

• NYC Mayor’s Office of Immigrant Affairs (MOIA) — convener, monitors language access across 46+ city agencies

• Population FactFinder — DCP tool for detailed census and ACS population profiles by geography

• NYC Department of City Planning, Population Division — home of Population Map Viewer and the upcoming Language Explorer

• CLEAR Global — international NGO working on language data and voice technology in humanitarian contexts

• CLEAR Global Language Maps and Data — open datasets and maps showing which languages are spoken where

• TWB Voice — CLEAR Global’s crowdsourced voice data collection platform for low-resource languages

• Translators without Borders — CLEAR Global’s community of 100,000+ language volunteers

• Endangered Language Alliance — documented 700+ language varieties spoken across New York City

• Local Law 30 Language Access Plans — citywide language access requirements and the 10 designated languages

The 130-Year-Old Entity That Can Fix NYC’s Broadband Crisis

A post by Suzi Ragheb and Katherine Jin for the NYC Policy Forum argues that New York City already possesses an overlooked tool that could significantly expand broadband competition and reduce Internet costs: the Empire City Subway Company (ECS), a 130-year-old underground conduit network that controls all of the underground conduits beneath Manhattan and the Bronx.

The authors frame New York’s broadband landscape as “digitally redlined,” with most residents limited to a single dominant provider or, at best, a duopoly between Verizon and Spectrum. High prices and weak competition, they argue, disproportionately harm low-income communities, particularly in the Bronx, where nearly one in four families cannot afford home broadband. At a May 2026 press conference announcing a new internet access initiative, a Bronx resident reported a broadband bill exceeding $200 a month.

Empire City Subway Company

The post explains that ECS was established in 1891 to move overhead telegraph and telephone wires off the street and into shared underground conduits. Today, ECS controls all underground conduit infrastructure throughout Manhattan and the Bronx.

Although ECS is now a Verizon subsidiary, the City retains powerful contractual rights over the system, including revenue-sharing claims and — under the original 1891 franchise — the rights to demand maps and occupancy data, require fair and impartial access for all tenants, and purchase the system outright.

The authors argue that this infrastructure could dramatically lower barriers for smaller fiber ISPs by allowing them to lease conduit space instead of incurring the enormous costs of street excavation, permitting, and new construction. In theory, this would let competitive providers reach residential customers at a fraction of the cost of building from scratch.

Mismanagement and Opacity

A major focus of the article is the lack of public accountability surrounding ECS operations.

The authors cite a 2010 audit by then-Comptroller John Liu, which found that ECS was undercounting profits — thereby reducing its required revenue sharing with the City — and failing to manage and reinvest in its network. The audit’s figures are striking: in 2008, ECS built 1,026 new conduits, of which 277 went to its parent Verizon and only 28 to other vendors, leaving 721 vacant. In 2007, competing vendors received just 4 of 1,484 newly built conduits.

The post notes that large amounts of potentially valuable public infrastructure remained locked away despite the city’s ongoing broadband affordability crisis.

The article also recounts efforts by Harvard law professor Susan Crawford, who in 2014 filed a Freedom of Information Law request seeking ECS’s franchise agreement, occupancy data, and financial records. According to the post, the City’s response was heavily redacted, with tenant information withheld as “trade secrets” on the grounds that disclosure would harm Verizon’s competitive position. Crawford sued the Office of Technology and Innovation (OTI), leading to a 2017 court ruling that pushed the City toward greater transparency around ECS assets.

Despite that ruling, the authors contend that Verizon continues to control access to the conduit system with minimal effective public oversight.

Policy Proposals

The post outlines several actions the City could take immediately under existing contractual authority, without new legislation.

The authors propose:

• A full public audit of ECS conduit capacity in the Bronx and Manhattan, with published occupancy data

• An open-access conduit leasing program with transparent, non-discriminatory pricing for competing ISPs

• Revenue-sharing terms tied to measurable competition outcomes (e.g., number of active conduit leases, pricing benchmarks, broadband adoption in underserved neighborhoods)

• Potential municipal acquisition of the ECS conduit system if Verizon refuses to open it to competitors

They argue that the City’s original franchise agreement already grants the authority to demand occupancy data, require fair tenant access, and purchase the system outright — all under OTI’s existing administrative authority.

Open Access and Competition

The central policy argument is that broadband competition is constrained less by technology than by physical access to infrastructure.

By opening conduit access to smaller providers, the City could foster greater ISP competition, potentially lowering prices and improving service quality. The authors emphasize the Bronx — the borough with the city’s worst broadband access — while noting that ECS covers only Manhattan and the Bronx, leaving other underserved boroughs out for now. They present ECS as a concrete, available, and underutilized policy lever rather than a speculative future project.

Broader Implications

The post situates the ECS issue within a larger national debate about municipal broadband and public infrastructure control.

The authors note that many municipalities across the country — and elsewhere in New York State — have built their own broadband networks, and they reference New York City’s own Neighborhood Internet pilot, a publicly owned network operated by the Department of Housing Preservation and Development and the New York Public Library. They frame ECS reform as aligned with the Mamdani administration’s broader push on internet access.

The piece concludes that New York’s broadband problems are not inevitable technological limitations but policy failures — addressable through infrastructure already embedded beneath the city’s streets.

Authors

Suzi Ragheb and Katherine Jin are leads for the Internet for All campaign, a grassroots, volunteer-driven initiative advocating expanded affordable Internet access for all New Yorkers.

COMMENTARY

• Ragheb & Jin, “New York City Should Stop Paying Corporations to Widen the Digital Divide” (Tech Policy Press) — the authors’ fuller April 2026 case for publicly owned broadband

• Susan Crawford, “I’m Suing New York City to Loosen Verizon’s Iron Grip” — the Harvard law professor’s account of her FOIL battle, central to the post’s transparency argument

• Courthouse News, “Secrecy Blamed for Slowing NYC Internet” — reports the 2010 Liu audit’s conduit figures and Crawford’s 2015 suit

• Crain’s New York Business, “Race is on to bring broadband to outer boroughs” — industry view; an ISP cites ECS conduit at ~$15/foot versus $200–400 to trench

• ILSR / MuniNetworks, “New York City Conduit Jam Packed” — civil-society analysis of conduit congestion and Verizon’s stewardship

• Bronx Times on the Neighborhood Internet expansion — reports roughly one in three Bronx households lack a computer; 113,000+ rely solely on a smartphone

• Yale MFIA, Crawford v. NYC DoITT case page — primary record of the FOIL litigation (2012 and 2014 requests; ECS, AT&T, Time Warner, RCN intervened)

RESOURCES

• “The 130-Year-Old Entity That Can Fix NYC’s Broadband Crisis” — the NYC Policy Forum post by Suzi Ragheb and Katherine Jin

• 2010 Comptroller audit of Empire City Subway (FP08-103A) — the Liu-era audit finding understated profits and unmanaged conduit build-out

• Empire City Subway Company — the Verizon subsidiary that controls underground conduit across Manhattan and the Bronx

• Internet for All — the authors’ grassroots campaign for open-access, publicly owned NYC broadband

• NYC Office of Technology and Innovation (OTI) — administers the ECS franchise and holds the relevant contractual authority

• Mayor’s Office release on the Neighborhood Internet expansion (May 2026) — $2M federal expansion of the City–NYPL broadband program

• ILSR Community Network Map — national survey of municipally owned broadband the authors invoke

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speaker: Shiva Muthiah - Designer and Developer;
Moderator: Shakil Assi - Mayor’s Office for Economic Opportunity

Introduction to PriceWise and Food Pricing Transparency

Shiva Muthiah introduced PriceWise as a community-built grocery price database designed to help budget-conscious New Yorkers make more informed food purchasing decisions. He described himself as a designer, developer, and photographer interested in making data more legible, accessible, and actionable, particularly for nonprofits and public-interest projects.

Muthiah explained that PriceWise emerged from his own personal experience after losing his job in 2025. He and his partner found it difficult to compare grocery prices across stores and determine where to shop most efficiently. This led him to initially build a private comparison tool, which later evolved into a broader community-oriented project.

He framed PriceWise as a response to growing food insecurity in New York City, citing several pressures:

• Roughly one in four New Yorkers live in poverty

• Many households spend up to 70% of income on food

• Food costs have risen by over 50% during the past decade

• SNAP benefits increasingly fail to fully bridge affordability gaps

Muthiah argued that food affordability problems are intensified by opaque pricing systems and algorithmic pricing practices used by large retailers.

Algorithmic Pricing and Information Asymmetry

A major theme of the presentation involved the lack of transparency in modern grocery pricing systems.

Muthiah discussed how large retailers increasingly use:

• Dynamic pricing systems

• E-Ink shelf labels

• Data brokers

• Machine learning models

• Loyalty program data

• Consumer browsing histories

to optimize prices in real time.

He explained that retailers can now rapidly adjust prices across entire stores within seconds and combine external data sources with customer purchasing behavior to personalize or optimize pricing strategies.

This creates what he described as a major “information asymmetry” between retailers and consumers. In theory, efficient markets assume consumers make rational choices using perfect information, but in reality consumers face:

• Limited information

• Limited time

• Limited financial flexibility

PriceWise was designed specifically to address the information gap by creating a bottom-up, community-generated grocery pricing database.

Muthiah contrasted these opaque systems with the Park Slope Food Coop, which openly discloses that it applies a 25% markup over wholesale prices, unlike many retailers that may apply markups between 30% and 80%.

Discussion with Participants on Desired Grocery Data

Muthiah paused to ask attendees how they would use open grocery pricing data if it were widely available citywide. Participants suggested many possible uses and related datasets, including:

• Comparing store prices more effectively

• Understanding which stores accept SNAP/EBT

• Identifying local versus imported products

• Supporting small “mom-and-pop” stores

• Understanding food transportation distances

• Tracking tariff impacts

• Finding stores with specific community features

The discussion reinforced Muthiah’s argument that consumers currently lack accessible, consolidated information needed to make informed food purchasing decisions.

Demonstration of the PriceWise Web Application

Muthiah then demonstrated the live PriceWise web application, available at “pricewise.nyc.” He emphasized that the platform is:

• Free to use

• Browser-based

• Accessible without downloading an app

• Anonymous, requiring no account creation

He demonstrated searching for grocery items while planning a Thai curry meal. Using mushrooms and coconut milk as examples, he showed how PriceWise allows users to:

• View price ranges for products

• Compare average and median prices

• Identify the cheapest nearby stores

• Evaluate whether a current in-store price is reasonable

The app displayed item prices connected to specific stores and neighborhoods, enabling users to make decisions about where to shop.

Muthiah also previewed future features that were still under development, including:

• Neighborhood-level filtering

• Personalized shopping lists

• Budget estimation tools

• Location-aware price comparisons

Community-Generated Receipt Data Collection

Muthiah explained that all pricing data inside PriceWise comes from community-contributed receipt photographs. Users simply photograph receipts with their phones and upload them to the platform.

He demonstrated the upload workflow:

1. Photograph a receipt

2. Select the store

3. Upload the receipt

4. Allow the system to process the image automatically

The backend processing pipeline then:

• Extracts text using OCR

• Structures receipt data using Gemini large language models

• Normalizes inconsistent item names

• Connects products with store and neighborhood datasets

• Adds the structured information into the searchable database

Muthiah emphasized that receipt normalization is especially important because stores use inconsistent naming conventions for identical products. Examples included differing descriptions for the same milk products or abbreviated item names on receipts.

The platform then presents extracted data alongside receipt images so users can verify or manually correct the OCR results before finalizing uploads.

Technical Architecture and Open Data Sources

Muthiah explained that PriceWise uses relatively lightweight and accessible open web technologies, including:

• HTML

• CSS

• JavaScript

• Flask

• SQLite

He credited NYC Open Data for supplying key datasets used by the application, particularly:

• Grocery store location datasets

• Neighborhood datasets

• GeoSearch APIs for mapping stores to neighborhoods

Later in the Q&A, he clarified that the project also relies heavily on the USDA Retail Food Stores dataset, which provides statewide grocery store information.

“The Price Is Wise” Interactive Quiz

Muthiah introduced a small experimental game called “The Price Is Wise,” designed to test participants’ intuition about grocery prices.

Participants answered questions involving:

• Matching food items with prices

• Comparing relative grocery costs

• Guessing unusually expensive items

One question revealed that milk — specifically a multi-pack milk product — was the single most expensive purchase recorded in the PriceWise database at that time, surprising both presenters and participants.

The game also illustrated how difficult it can be for consumers to estimate fair grocery prices accurately.

Short-Term Challenges: Bootstrapping and Messy Data

Muthiah described several major short-term challenges facing the project.

The first was a classic “bootstrapping” problem:

• The platform needs enough data to become useful

• People are less likely to contribute data if the database remains sparse

To address this, Muthiah has been conducting grassroots outreach through:

• Open Data Week events

• Friends and neighbors

• Potential future outreach at farmers markets

• Community organizations

• Mutual aid groups

• Local elected officials

Another major challenge involves messy receipt data. He explained that receipts vary dramatically in:

• Layout

• Formatting

• Naming conventions

• Quantity notation

• Pricing structures

As a result, the system requires both automated normalization techniques and manual curation strategies to unify comparable products.

Long-Term Challenges: AI, Ethics, and Sustainability

Muthiah devoted substantial time to broader long-term concerns surrounding AI systems, transparency, and sustainability.

He noted that AI systems are inherently non-deterministic, meaning identical receipts can sometimes produce slightly different extraction results.

He also raised ethical concerns involving:

• Intellectual property issues in AI training

• Energy consumption

• Dependence on large commercial AI systems

Muthiah suggested that future versions of PriceWise might eventually transition toward smaller, more specialized models trained specifically for receipt analysis.

Another concern involved ownership and governance of community-created data. He emphasized that because the database is built collectively by users, safeguards are needed to prevent exploitation or privatization of community-generated pricing data.

Muthiah also discussed the importance of transparency in both:

• The software itself

• The decision-making logic embedded within the system

He noted that open-sourcing code alone does not guarantee accessibility or understanding for ordinary users.

Finally, he discussed sustainability challenges, including:

• Hosting costs

• Domain expenses

• AI API costs

• Time investment

At the time of the presentation, hosting costs were approximately $12–20 per month, while Gemini API usage had totaled roughly $11 after processing hundreds of receipts. However, he emphasized that the project represented hundreds of unpaid hours of personal labor.

Comparison with Similar Grocery Data Projects

Muthiah reviewed several other grocery-pricing and food-data initiatives for comparison.

These included:

• Savvy Prices

• Basket

• Fetch

• Open Prices

• Open Food Facts

• Matpriskollen (Sweden)

He explained that many competing systems focus primarily on scraping data from major retail chains, which excludes small independent grocery stores that are common throughout New York City.

The Swedish platform Matpriskollen particularly inspired him because Swedish regulators encouraged large retailers to share pricing data more openly during periods of concern over food inflation. Muthiah suggested that a similar future could potentially emerge in the United States.

Future Roadmap for PriceWise

Muthiah outlined a three-phase vision for PriceWise.

Phase One:

• Building core infrastructure

• Receipt ingestion

• Data normalization

• Database creation

Phase Two:

• Personalized shopping lists

• Neighborhood-specific recommendations

• Budget forecasting

• More actionable user tools

Phase Three:

• Advanced visualizations

• Geographic price comparisons

• Historical price tracking

• SMS-based access

• EBT-focused accessibility features

He emphasized that future development would require deeper collaboration with food policy experts, community organizations, and researchers working on food access and food equity issues.

Questions About Store Distance, Barcodes, and Dynamic Pricing

During the Q&A session, participants asked numerous technical and usability questions.

Muthiah confirmed that future versions could include distance-based shopping optimization, allowing users to identify the cheapest prices within a specified walking radius.

Questions also addressed:

• Handling extremely long receipts

• Simplifying the interface for less technical users

• Showing upload dates for pricing data

• Expanding the system to other cities such as Boston

Participants proposed using UPC barcodes to standardize products across stores. Muthiah acknowledged this as a promising future direction and noted that Open Food Facts already uses similar barcode-based approaches.

He also acknowledged that the current system does not yet fully account for:

• Sale pricing

• Temporary discounts

• Coupons

• Dynamic price fluctuations

These remain active development challenges.

Discussion on Public Policy and Economic Research Applications

Participants suggested that PriceWise could potentially support broader economic research, including price-index calculations used by the Bureau of Labor Statistics.

Muthiah responded that he had studied some inflation-tracking methodologies and was interested in potentially integrating benchmark pricing information, such as Park Slope Food Coop wholesale markup data, into future analyses.

Closing Discussion on Community Outreach and Scaling

Toward the end of the session, Muthiah reflected on future outreach plans and technical scalability.

He explained that the system had already been tested with approximately 3,000–4,000 purchases and was functioning reliably at that scale.

Future outreach plans include:

• Farmers market demonstrations

• Partnerships with mutual aid organizations

• Engagement with community boards

• Collaboration with block associations

The session concluded with broader reflections on the importance of making pricing data more legible, transparent, and accessible for ordinary residents navigating rising food costs and increasingly opaque retail systems.

RESOURCES

• PriceWise — community-built grocery price database for NYC, presented by Shiva Muthiah (currently in beta)

• Shiva Muthiah — designer, developer, and photographer who built PriceWise

• Retail Food Stores dataset — NYS open dataset of licensed grocery stores, published by the Department of Agriculture and Markets

• NYC GeoSearch API — address-to-neighborhood geocoding service used to map stores to neighborhoods

• Open Prices — crowdsourced open database of food prices, part of Open Food Facts

• Open Food Facts — open barcode-based food product database referenced for item identification

• Matpriskollen — Swedish grocery price comparison service cited as an inspiring model for data-sharing

• Savvi — Vancouver-focused grocery price comparison tool

• Park Slope Food Coop — Brooklyn cooperative noted for its transparent flat 25% markup on wholesale prices

• NYC Open Data — the city’s public data portal, source of the grocery store and neighborhood datasets

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speaker: Ora Kemp - Senior Policy Advisor, NYC Mayor’s Office of Food Policy

Introduction to the Supply Gap Analysis and Open Data Resources

Ora Kemp opened the session by thanking the Open Data Week organizers and noting that this was her second year presenting on New York City’s emergency food “supply gap” analysis. She described the work as a subject she was deeply passionate about and emphasized that the session would focus on the methodology behind the analysis and how the city uses it to make funding and operational decisions.

Kemp introduced several public resources connected to the project, including:

• The NYC Open Data supply gap dataset

• An interactive Tableau-based supply gap map

• The Mayor’s Office of Food Policy supply gap information page

• Community Food Connection (CFC) program resources

She explained that the dataset is updated annually and includes multiple years of consecutive tracking data. The Tableau visualization allows users to explore neighborhood-level food insecurity and emergency food supply conditions interactively.

Kemp also acknowledged the contributions of Adam Santos and Claire Reynolds from the Mayor’s Office for Economic Opportunity, describing them as major collaborators in earlier iterations of the analysis before responsibility transitioned to the Department of Social Services data team.

She framed the session around understanding how data insights guide:

• Public funding allocation

• Philanthropic investment

• Organizational strategies

• Neighborhood-level collaboration

The goal, she explained, was to understand how emergency food need is defined, measured, and operationalized into city policy.

Community Food Connection Program Overview

Kemp introduced the Community Food Connection (CFC) program, operated under the NYC Department of Social Services. The program supports food pantries, soup kitchens, and emergency food providers through direct city funding.

She explained that CFC primarily operates through two mechanisms:

1. Direct food purchasing support

2. Capacity-building support

The purchasing component provides emergency food organizations with city-funded purchasing power that can be used to obtain approved food products. Kemp noted that the approved product list can adapt over time in response to community needs.

The second component supports organizational infrastructure and scalability, including:

• Refrigeration and storage

• Transportation

• Technology

• Administrative capacity

• Staffing support

Kemp emphasized that rising food prices and broader cost-of-living pressures continue to increase demand on emergency food providers, making organizational capacity investment increasingly important.

Neighborhood Tabulation Areas and TRIE Priority Neighborhoods

Kemp explained that the city analyzes food need using Neighborhood Tabulation Areas (NTAs), which divide New York City into 197 residential neighborhoods. These smaller geographic units allow more granular analysis of neighborhood-level conditions.

She also described the Taskforce on Racial Inclusion and Equity (TRIE) neighborhood designation system, created during the COVID-19 pandemic to identify neighborhoods facing overlapping challenges such as:

• High COVID infection rates

• Healthcare access barriers

• Socioeconomic instability

• Structural inequities

The city uses both NTA-level data and TRIE priority designations to guide emergency food funding and policy decisions.

Scale of the Emergency Food Network

Kemp presented 2025 metrics for the Community Food Connection program:

• 703 food access organizations received direct support

• Over 47 million pounds of food were distributed

• More than 37 million pantry visits occurred

She clarified that visit counts represent repeated visits rather than unique individuals.

The 47 million pounds distributed through CFC-funded organizations accounted for approximately 24% closure of the citywide emergency food supply gap. Kemp emphasized that CFC-funded organizations represent only about 20% of the total emergency food distributed citywide. The remaining 80% comes from organizations outside the CFC funding structure, including churches, independent pantries, and community organizations.

Across the broader emergency food network, approximately 241 million pounds of food were distributed citywide. Kemp repeatedly stressed that even this large dataset is not fully comprehensive because not every emergency food provider reports into the city’s systems.

Universal School Meals and Food Access

Responding to a participant question about hunger and schools, Kemp explained that New York City operates under the federal Community Eligibility Provision, which provides free breakfast and lunch to all public school students regardless of income. She noted that the policy was later adopted statewide in New York.

Coverage and Reach of the CFC Network

Kemp explained that the CFC network operates across 170 of the city’s 197 NTAs. While this represents broad coverage, some neighborhoods still lack direct CFC-supported emergency food providers.

She stated that approximately 1.5 million New Yorkers are considered food insecure, and CFC pantry locations serve neighborhoods containing roughly 92% of that population.

Kemp emphasized the importance of collaboration across the broader emergency food network to prevent geographic service gaps.

Defining the Supply Gap

Kemp then explained the core logic of the supply gap analysis:

Demand for emergency food minus supply of emergency food equals the supply gap.

If need exceeds supply, a deficit exists. If supply exceeds need, there is a surplus.

The annual analysis consists of two major components:

1. Calculating the raw supply gap

2. Applying a prioritization process that incorporates additional social vulnerability factors

She explained that the city uses Feeding America’s “Map the Meal Gap” dataset as the foundation for estimating food insecurity. These estimates are derived from:

• American Community Survey data

• Household Food Security Survey data

• Federal-level food insecurity modeling

Kemp noted that emergency food demand estimates are converted into pounds of food to create a standardized measurement system across organizations and programs.

Threats to Federal Food Security Data Collection

Kemp raised concerns about the elimination of the Household Food Security Survey under HR 1, explaining that the removal of this survey threatens the foundational data used for national food insecurity estimates.

She said that city and state partners are now exploring alternative methodologies for measuring household food insecurity in the absence of the federal survey. Kemp suggested this methodological transition would likely become a major topic at future Open Data Week sessions.

FeedNYC Platform and Supply Tracking

Kemp explained that emergency food supply data comes from the FeedNYC platform, through which providers report:

• Pounds of food distributed

• Meal counts

• Pantry visits and contacts

Each participating Emergency Food Relief Organization (EFRO) receives a unique identifier, allowing the city to track food distribution geographically.

The analysis incorporates a half-mile catchment radius around each EFRO location to account for the distance residents are likely to travel to access food resources.

Kemp explained that the city estimates roughly 321 million pounds of emergency food would be required annually to fully meet food insecurity needs citywide.

Mapping Need, Supply, and Surplus

Kemp walked through visualizations showing:

• Neighborhood-level food insecurity need

• Locations of EFROs

• Geographic food supply concentrations

• Supply deficits and surpluses

The analysis showed especially high need in:

• The Bronx

• Upper Manhattan

• Southwest Brooklyn

• Northeast Brooklyn

• Jackson Heights and Elmhurst in Queens

Between 2024 and 2025, approximately 1,084 organizations reported into FeedNYC, supplying 242.5 million pounds of food.

Kemp explained that neighborhoods shown in yellow on the supply gap map represented areas with unmet need, while darker blue areas represented relative surplus.

She cautioned against simplistic redistribution strategies that remove supply from surplus neighborhoods to serve deficit neighborhoods, noting that surplus areas may still require substantial support to maintain stability.

Growing Hunger and the Food Gap Metric

One of the session’s most important metrics involved pounds of unmet food need per food insecure resident.

Kemp explained:

• The previous year’s average gap was 47 pounds per person

• The current year’s gap rose to 52 pounds per person

This increase demonstrated that hunger is growing faster than the city’s emergency food response capacity.

She argued that the city therefore faces two simultaneous policy challenges:

• Increasing emergency food supply

• Addressing underlying economic causes of food insecurity

Neighborhood Prioritization Rankings

Kemp described the neighborhood prioritization ranking system used to allocate funding. Every NTA receives a rank from 1–197, with lower numbers representing neighborhoods requiring the greatest investment.

The ranking incorporates several weighted factors beyond raw supply gap numbers:

• Supply gap above city average

• TRIE designation

• Unemployment rates

• SNAP eligibility

• Vulnerable populations

• Youth populations

• Older adults

• Foreign-born non-citizens

• Veterans

Kemp explained that these additional variables help account for real-world conditions not fully captured in the underlying federal datasets.

Top Priority Neighborhoods and SNAP Policy Impacts

Kemp presented examples showing that neighborhoods with slightly smaller raw food gaps could still rank higher if they faced additional vulnerabilities such as unemployment or concentrated vulnerable populations.

She also discussed how recent SNAP benefit cuts influenced city emergency food planning. When SNAP reductions occurred, the city redirected additional emergency food resources toward neighborhoods with high SNAP dependency.

Tableau Visualization Demonstration

Kemp demonstrated the interactive Tableau map using her own neighborhood, Mott Haven in the South Bronx. She showed how users can click into neighborhoods to view:

• Weighted neighborhood scores

• Food insecurity rankings

• Unemployment rankings

• Supply gap metrics

She encouraged attendees to explore their own neighborhoods and use the data for community advocacy, city council engagement, and organizational planning.

Mobile Pantries and Alternative Distribution Models

Responding to questions about mobile pantries, Kemp explained that mobile pantry distributions are currently difficult to map precisely because they operate across multiple locations.

At present, most mobile pantry distributions are attributed to the organization’s primary location. She acknowledged that this creates limitations in accurately representing neighborhood-level food access.

Kemp discussed broader experimentation with:

• Pop-up food distributions

• Mobile operations

• Refrigerated locker systems

• Direct delivery systems

• Last-mile food delivery models

Case Studies in Persistent Need and Service Disruption

Kemp presented several case studies demonstrating how the data informs policy decisions.

One example involved seven neighborhoods that remained in the top 20 highest-need areas across three consecutive years. She argued that persistent need indicates broader structural challenges beyond food distribution itself, including:

• Housing instability

• Healthcare costs

• Transportation barriers

• Broader cost-of-living pressures

Another case study showed how the closure or relocation of a single high-volume pantry could dramatically alter neighborhood rankings. Kemp described a case where Corona Park shifted from rank 194 to rank 23 after one major food distribution site closed.

The city responded by exploring:

• Mobile pantry deployment

• Pop-up distributions

• Emergency resource partnerships

• Rapid collaboration with other providers

Gentrification, Displacement, and Real-Time Data Challenges

Kemp acknowledged that the city’s core datasets are inherently delayed because they rely on census-based information that may lag real-world neighborhood change by roughly two years.

To compensate, the city increasingly relies on adaptive prioritization factors and is exploring additional metrics such as:

• Deviation from area median income

• Real-time indicators of displacement

• Dynamic neighborhood migration patterns

Kemp explained that mobile food operations have become especially important because they can adapt more rapidly to population displacement patterns than fixed brick-and-mortar pantries.

Integrated Social Policy and Food Security

Toward the end of the session, Kemp emphasized that food insecurity cannot be understood in isolation from broader economic pressures. She argued that by the time families struggle to access food, they have often already experienced:

• Housing instability

• Transportation insecurity

• Healthcare costs

• Income volatility

She described food as one of the final flexible household expenses after rent and other fixed costs are paid. Because of this, emergency food demand often reflects deeper structural instability throughout the city economy.

Kemp stressed the need for integrated approaches that connect food policy with housing, transportation, healthcare, and social safety net systems.

Food Waste, Surplus Redistribution, and donateNYC Food

The session concluded with discussion of food waste and surplus redistribution. Kemp noted that New York City continues to face major challenges involving food surplus and waste even while hunger increases.

She highlighted organizations working on food rescue and redistribution and discussed the donateNYC Food platform, which aims to connect surplus food providers with organizations capable of redistributing food rapidly before spoilage occurs.

Kemp concluded by emphasizing the importance of strengthening citywide systems that can connect surplus food resources with communities experiencing unmet need while reducing food waste throughout the emergency food ecosystem.

RESOURCES

• Emergency Food Supply Gap dataset — multi-year NYC Open Data dataset by Neighborhood Tabulation Area

• Supply Gap Analysis — Mayor’s Office of Food Policy page with the interactive Tableau Supply Gap Map

• Supply Gap Analysis FAQ — paper-format Q&A on methodology and how the tool is used

• Community Food Connection (CFC) — the NYC emergency food program funded by the supply gap analysis

• Community Food Connection Quarterly Report — NYC Open Data dataset tracking the CFC program

• Feeding America Map the Meal Gap — the federal study supplying NYC’s emergency food demand estimates

• FeedNYC — citywide reporting platform for emergency food providers, the source of supply figures

• NYC Mayor’s Office of Food Policy — the office leading the supply gap analysis and CFC program

• NYC Mayor’s Office for Economic Opportunity — developed the supply gap analysis methodology in prior years

• donateNYC — DSNY’s food donation and reuse platform discussed for reducing food waste

VIDEO | AUDIO | RECAP EN / ES / FR | INFO | INDEX

Speakers: Matt Gold - Associate Professor of English and Digital Humanities, The Graduate Center, CUNY; Joel Christensen - Provost and Senior Vice President for Academic Affairs, The Graduate Center, CUNY; Lisa Gelobter - NYC CTO / Commissioner, Office of Technology & Innovation (OTI); Tayyab Walker - Associate Commissioner, Office of Technology & Innovation (OTI); Andrew Kittredge - Director, Civic Innovation Lab, BetaNYC; Rahul Bhargava - Assistant Professor, Northeastern University

Moderator: Matt Gold - Associate Professor of English and Digital Humanities, The Graduate Center, CUNY

Opening Remarks and Festival Context

Matt Gold welcomed attendees to the kickoff event for NYC Open Data Week 2026, hosted at the CUNY Graduate Center in partnership with the NYC Office of Technology and Innovation (OTI). He highlighted the festival’s 10th anniversary and thanked organizers Claudia Berger and Zachary Federer from OTI. Gold also acknowledged sponsorship from the Graduate Center’s MA Program in Digital Humanities and MS Program in Data Analysis and Visualization.

Gold emphasized the Graduate Center’s commitment to creative and socially engaged uses of data, referencing a textile data quilt by alumna Paula Fleischer documenting the experiences of Argentine mothers whose children disappeared during the country’s “Dirty War.” He framed the keynote as part of a broader effort to connect data practices with public storytelling and civic engagement.

Joel Christensen on Public Knowledge and Digital Scholarship

Joel Christensen situated Open Data Week within the Graduate Center’s public mission as part of the largest urban university system in the United States. He described the institution’s role as preserving, transmitting, and creating knowledge “for the public good,” emphasizing that open data and public-interest technology are central to ensuring communities benefit from modern computing and AI rather than merely being shaped by them.

Christensen highlighted the Graduate Center’s expanding programs in digital humanities, data analysis, data science, computational linguistics, and critical AI studies. He stressed the importance of equipping students and city residents with tools to critically engage with technology, arguing that universities must help communities become “agents and subjects” rather than “objects and targets” of digital systems.

Lisa Gelobter on Open Data and Civic Impact

Lisa Gelobter, newly appointed as NYC CTO and OTI Commissioner, reflected on her earlier work at the U.S. Department of Education during the Obama administration, particularly the creation of the College Scorecard initiative. She described how opening educational data through APIs enabled broader public access and third-party innovation beyond government websites.

Gelobter argued that open data’s greatest power lies not merely in publication but in accessibility and distribution. She explained that the College Scorecard data release reshaped conversations around higher education by emphasizing affordability, access, and outcomes rather than prestige alone. She noted that the initiative was credited with improving college graduation rates nationwide by approximately 1.5 percentage points.

She connected those experiences to NYC Open Data, emphasizing its role in serving 8.5 million New Yorkers and positioning public data infrastructure as essential civic infrastructure. She also praised the festival’s broad programming, including data art exhibitions, documentary screenings, and conferences.

Tayyab Walker on Open Data as Civic Infrastructure

Tayyab Walker described the Office of Data Analytics as the city’s central data office, organized around four pillars: actionable data science, practical data governance, responsible data sharing, and robust civic engagement.

Walker emphasized that Open Data Week emerged from the belief that public data should not merely be published but should help residents understand government operations and hold institutions accountable. He described the festival’s growth over ten years into a large civic ecosystem involving academics, data practitioners, artists, advocates, and residents.

He highlighted the 2026 festival’s scale:

• More than 80 events

• Over 20 city agency-hosted sessions

• More than 40 events hosted by academic institutions

• 21 events hosted by CUNY

Walker also emphasized the festival’s creative direction, which included data art, documentary screenings, weaving, and even sock knitting, arguing that “data is more than spreadsheets.” He promoted the Data Through Design exhibition at BRIC in Brooklyn and thanked BetaNYC and community organizers for sustaining the initiative.

Andrew Kittredge on Civic Community and Participation

Andrew Kittredge of BetaNYC framed Open Data Week as a community-building exercise rooted in connection, collaboration, and civic participation. He stressed that the festival’s value lies not merely in the number of sessions but in the relationships and networks it creates.

He argued that in an era of fragmentation and isolation, events like Open Data Week help people feel welcomed, connected, and able to participate in shaping civic technology and public-interest innovation. He encouraged attendees to participate in the School of Data conference and the inaugural “Unschool of Data.”

Rahul Bhargava’s Framework: Community Data

Rahul Bhargava opened by positioning himself as an educator, designer, artist, and “recovering computer scientist” whose work explores creative approaches to data storytelling, civic participation, and community empowerment.

He argued that traditional data tools — spreadsheets, charts, and dashboards — are insufficient for helping broader communities meaningfully engage with information. While those tools evolved in scientific and business contexts, he explained that data now permeates civic life, journalism, libraries, museums, and community organizing.

Bhargava’s central argument was that society needs “a bigger toolbox” for engaging people with data, particularly approaches that:

• Invite impactful participation

• Open doors to layered stories

• Build “mirrors, not windows”

He argued that conventional data visualization often positions audiences as detached observers looking through a “window” at others, whereas community-centered approaches should instead function as “mirrors” that help people see themselves and their communities reflected in data.

Data Sculpture and Embodied Storytelling

Bhargava introduced a series of experimental examples demonstrating alternative forms of data communication.

The first example was a data sculpture created by Colombian artist Jose Duarte representing changes in marine protected areas connected to the UN Sustainable Development Goals. Rather than presenting the information through charts, the data was encoded in physical sculptural form to create an emotional and tactile connection to environmental progress.

Bhargava contrasted optimistic statistical narratives with a quote from Uruguayan writer Eduardo Galeano: “In Central America, the more wretched and desperate the people, the more the statistics smiled and laughed.” He used the quote to illustrate how numerical systems can become disconnected from lived experience.

Climate Sonification Exercise

Bhargava then led the audience through an interactive “sonification” exercise translating national carbon emissions data into rhythm and percussion.

Audience members collectively represented Brazil, India, and the United States through different rhythmic patterns corresponding first to total emissions and then to per-capita emissions. By changing the sonic balance between countries, the audience experienced how normalization changes the interpretation of climate data.

Bhargava argued that performing data physically and collectively activates different forms of understanding than reading charts alone. He explained that embodied participation:

• Makes data memorable

• Creates emotional investment

• Encourages collaborative interpretation

• Engages people who might otherwise avoid quantitative material

He connected the exercise to broader research on music, embodiment, and civic engagement, noting that audiences often leave sonification experiences feeling more emotionally connected to climate issues than after reading conventional reports.

Data Theater and Rehumanizing Statistics

Bhargava described ongoing collaborations through the Civic Data Theater project, which combines theater, deliberation, and data visualization to facilitate community conversations around issues like climate adaptation and urban green space.

He referenced theatrical traditions inspired by Augusto Boal and participatory performance practices, explaining how theater can create empathy and collective understanding around complex issues.

One example involved a workshop based on migration data where participants carried balloons representing migrants crossing borders. As balloons were popped unexpectedly during the performance, participants experienced anxiety and vulnerability connected to the realities represented in the data.

Bhargava argued that these approaches “rehumanize” data that might otherwise feel abstract or dehumanizing.

Edible Data and Air Quality Brownies

In one of the keynote’s most memorable demonstrations, Bhargava distributed brownies baked with varying levels of salt corresponding to New York City air quality levels during the 2023 Canadian wildfire smoke events.

Audience volunteers tasted brownies representing:

• A normal low-pollution day

• A day at the federal air-quality threshold

• An extremely hazardous pollution day

The increasing saltiness translated environmental exposure into a sensory experience. Bhargava argued that multisensory methods create “doorways” into public understanding by making abstract metrics physically tangible.

He linked the exercise to museum exhibit design principles, where attracting curiosity can open opportunities for deeper learning and civic action.

Food Security Table Sculpture

Bhargava also presented a large sculptural table constructed from 1,659 pieces of cutlery, representing the number of Massachusetts households applying daily for SNAP food assistance during the peak of the COVID-19 pandemic.

Created with artist Emily Bhargava, the piece was exhibited at:

• Farmer’s markets

• Galleries

• Government spaces

• Community events

Bhargava explained that the sculpture’s spectacle attracted public attention, allowing organizers to direct people toward food assistance programs, local organizations, and policy advocacy efforts. He connected the work to historical traditions of activist infographics and abolitionist visual campaigns.

Wearable Data and Community Mirrors

Bhargava then introduced a wearable “data quilt” jacket created by Claudia Berger using Bengali textile traditions, sashiko stitching, and American quilt patterns to encode urban environmental data.

He argued that wearable and tactile data objects help communities build “mirrors” reflecting their own experiences rather than detached visualizations aimed at outsiders.

Bhargava described his broader “data mural” practice, in which communities collaboratively interpret local data, design visual narratives, and paint murals together. He emphasized that the collaborative process itself often becomes as important as the final artwork because it builds relationships among residents, nonprofits, artists, and city agencies.

Discussion on Trauma, Community Practice, and Participation

During the Q&A, audience members raised questions about trauma-informed facilitation, critical data studies, community sustainability, and institutional power.

Bhargava explained that he often works alongside trusted community organizations already embedded in local neighborhoods rather than entering communities independently. He emphasized the importance of trauma-informed methods borrowed from theater and participatory arts traditions.

When asked about critical perspectives on datafication and extractive quantification, Bhargava acknowledged those critiques and referenced collaborations with scholars such as Catherine D’Ignazio and Lauren Klein, particularly around feminist and anti-extractive approaches to data practices.

Data Literacy, Power, and Community Capacity

Bhargava reflected on his earlier work in “data literacy,” noting that over time he became uncomfortable with deficit-oriented assumptions that framed communities as lacking data skills.

Instead, he increasingly focused on recognizing and activating existing community assets — artists, musicians, storytellers, organizers, and local knowledge — as resources for engaging with information and civic power.

He stressed that data functions as “a language of power,” and that helping communities creatively engage with data can open pathways to political participation, advocacy, and institutional accountability.

Data Creation Versus Data Collection

In the final discussion, Bhargava challenged the idea that data simply exists waiting to be collected. He argued instead that communities actively create data through observation, documentation, and collective inquiry.

Using examples such as youth documenting tobacco advertising in corner stores, he emphasized that community groups can generate their own evidence and narratives rather than relying entirely on institutional datasets.

He concluded by encouraging technologists, researchers, and organizers to use data practices to help communities access power and shape the systems affecting their lives.

RESOURCES

• Community Data: Creative Approaches to Empowering People with Information — Rahul Bhargava’s book from Oxford University Press, the basis of the keynote

• Community Data book companion site — examples, reviews, and interviews accompanying the book

• Data Culture Group — Bhargava’s research group at Northeastern University’s College of Arts, Media and Design

• DataBasic — the self-service suite of data literacy tools and activities Bhargava built with Catherine D’Ignazio

• Data Theatre Collaborative — Northeastern’s Civic Data Theatre project, supported by the Mellon Foundation

• Data Feminism — the book by Catherine D’Ignazio and Lauren Klein referenced during the Q&A

• NYC Open Data Week — the annual festival this keynote kicked off, marking its 10th anniversary

• NYC Open Data — the city’s free public data portal at the center of the festival

• BetaNYC — civic organization and festival co-organizer, home of the Civic Innovation Lab

• Data Through Design — art collective and festival co-organizer behind the annual NYC Open Data art exhibition

New York City Comptroller Mark Levine’s office released a report on May 21, 2026, that does something most AI commentary avoids: it puts numbers on the uncertainty. Rather than predicting a single future, the Bureau of Budget models five distinct scenarios for how AI could reshape the city’s economy, jobs, and tax base through 2030 — and assigns each a probability. The conclusion is not a forecast but a fiscal posture: build the reserves now, because the city is dangerously underprepared for the downside cases.

A telling detail sits in the adoption data. By the official federal measure (Census Business Trends and Outlook Survey), New York State actually lags the national AI-adoption rate — 16.8 percent of establishments versus 19.8 percent nationally. But usage tells a different story. Anthropic’s Economic Index puts New York at roughly 9 percent of national Claude usage against a ~6 percent share of national employment, with adoption sharpest in finance, information, and professional services — precisely the sectors that drive the city’s wages and tax receipts. The report also notes Anthropic’s expected lease of 330 Hudson Street, one signal among several that AI firms are expanding their physical footprint in Manhattan even as white-collar hiring stalls elsewhere.

That stall is the report’s most concrete present-day finding. The city is in a “low-hire, low-fire” economy: net job creation has been moribund while layoffs stay historically low, breaking a correlation that held from 1975 to 2019. Entry-level white-collar roles look most exposed, and the effect on recent graduates is striking — for the first time on record, in the twelve months ending March 2026, NYC college graduates aged 22–27 faced a slightly higher unemployment rate (7.3 percent) than young adults without degrees (7.1 percent). Meanwhile the Manhattan office market is booming, with nearly 31 million square feet of new leasing in 2025.

The five scenarios. The office began with four macro scenarios from Moody’s Analytics, adapted them to NYC’s industry mix, and added a fifth — a more severe white-collar shock — trimming the baseline probability from 40 to 35 percent to make room for it:

• AI-Empowered Economy (35%) — the baseline. Modest productivity gains, steady job growth, tax revenue rising ~3.1 percent annually.

• AI Falls Flat (25%) — the investment boom fizzles, markets sell off ~35 percent, a brief recession costs NYC ~52,500 private-sector jobs over the year (a ~135,000 gap below baseline at the 2027Q3 trough) and a $3.4 billion FY2027 revenue hit, with cumulative losses near $9 billion.

• Job Replacement (20%) — faster adoption replaces routine cognitive work; ~96,000 fewer private-sector jobs than baseline by 2030, ~$5.5 billion cumulative revenue shortfall, though Wall Street stays strong.

• Productivity Boon (15%) — the optimistic case, akin to the late-1990s internet boom. Growth and wages accelerate; ~$8 billion more in cumulative tax revenue. The only scenario that beats baseline on every tax category.

• AI Shockwave (5%) — the tail risk. More than 110,000 jobs lost in 2027 alone, a peak gap of ~259,000 below baseline in early 2029, three straight years of declining Wall Street profits, and ~$14 billion in cumulative revenue losses through FY2030. Low probability, highest impact.

One important caveat the report flags itself: the scenarios were drawn before the war in Iran, and neither that conflict nor the subsequent run-up in asset values is factored in. Middle East geopolitics and energy costs sit as downside risks across all five paths.

The actual ask. Three of the five scenarios are negative, together carrying 50 percent of the probability. The report’s central recommendation is that the city stop “sleepwalking into the age of AI” and rebuild its fiscal cushion. By S&P’s measure, NYC had the second-lowest available reserves as a share of operating revenue among the ten largest U.S. cities in FY2025, ahead only of Chicago. The Comptroller endorses a rainy-day-fund target of 16 percent of annual tax revenues — about $13.5 billion against projected FY2026 tax revenues of $84.4 billion. The current cushion (Revenue Stabilization Fund plus Retiree Health Benefit Trust) holds just $7.2 billion, or 8.5 percent. Notably, $13.5 billion is roughly the size of the AI Shockwave’s projected revenue hit — the reserve is sized to absorb exactly the scenario the report fears most.

The framing is that uncertainty is a reason to prepare, not an excuse to wait. Whether AI proves to be the internet boom or a white-collar shock, the report argues, the city’s ability to protect services and support displaced workers will be decided now, by how large a buffer it chooses to build while revenues are still strong.

RESOURCES

• AI and New York City’s Fiscal Future — the full Office of the NYC Comptroller report, May 21, 2026

• Office of the New York City Comptroller — Mark Levine, Bureau of Budget

• Strengthening the City’s Rainy Day Fund — the April 2026 proposal behind the 16 percent target

• Anthropic Economic Index — the Claude-usage data showing NYC at ~9 percent of national use

• Moody’s Analytics — source of the four base macroeconomic scenarios

• Census Business Trends and Outlook Survey — the BTOS adoption measure (NY at 16.8 percent)

• Federal Reserve Bank of Atlanta — the ~750-CFO survey on AI investment and the workforce

• Federal Reserve Bank of St. Louis — study attributing ~40 percent of 2025 GDP growth to AI investment

• IEA: Energy and AI — data-center electricity demand projections cited as a constraint

• Ethan Mollick — Wharton professor quoted on uncertainty versus helplessness