NYC 311 Data Maps the Complaints That Set Neighborhoods Apart

New York City’s 311 records can show more than the problems residents share. By comparing complaint patterns across neighborhoods, Kieran Healy’s analysis surfaces the reports that are unusually common in each place, from lost property near nightlife districts to vehicle complaints elsewhere. The result is a map built to provoke recognition and amusement, with a method that also offers a useful lesson: complaint data describes what people report, not everything happening around them.

From 43 million service requests to neighborhood patterns

The underlying dataset contains roughly 43 million service requests logged between 2010 and 2025. New York City’s 311 phone line and website handle non-emergency requests, and the city publishes anonymized records through NYC OpenData. Healy packaged the records for analysis in R, using a large Parquet file queried with DuckDB.

That setup matters for the scale of the work. DuckDB is designed to query large datasets efficiently, which suits repeated analysis better than a workflow centered on updating individual records. The data can be accessed through R with dplyr-style queries, while the database engine handles the work against the stored file. Each record includes fields such as the date opened and closed, agency, complaint category, location, borough, and coordinates.

But a large dataset isn’t automatically a consistent one. Complaint categories have changed over time, and some were substantially reworked. A change in the city’s data management provider also affected measurement. The volume of reports can reflect changes in reporting behavior, too: a web form makes submitting a complaint quicker than placing a call. And some residents report problems far more often than others. Those quirks shape what any neighborhood comparison can reasonably claim.

Why count addresses instead of reports?

For the neighborhood analysis, Healy focused on the years 2022 to 2025, after harmonizing complaint categories across the dataset. That period contributes about 13.5 million records. He then grouped reports by New York City’s Neighborhood Tabulation Areas, or NTAs, a unit designed to sit between census tracts and the city’s larger community districts.

New York has 262 NTAs, 197 of them residential. The city’s NTA system provides a practical scale for comparing local patterns: its areas are larger than the city’s roughly 2,300 census tracts, yet smaller than its 59 community districts.

Rather than letting frequent reporters dominate the results, the analysis counts distinct addresses in each complaint category within each residential NTA. An address counts once for a category over the period, even if it generated multiple reports in that category. The resulting dataset has about 4.6 million address-category pairs, drawn from approximately one million unique addresses. The addresses don’t include identifying details such as house or apartment numbers. On average, each address appears in four or five complaint categories during the four-year window.

That choice shifts the question from “Where were the most reports filed?” to “Which types of complaint came from the greatest share of addresses in each area?” It reduces the influence of people who file repeatedly, while preserving the fact that one location might generate complaints across several different categories. It’s a useful adjustment for comparing neighborhood profiles, though it still measures reported concerns rather than independently verified incidents.

Finding what’s distinctive, not merely common

Many of the most frequent complaints are common across the city: heating or hot water problems, different kinds of noise, and illegal parking. Ranking each neighborhood by its raw complaint totals would mostly reproduce that shared list. To make local differences visible, Healy compares each category’s share of complaints citywide with its share within an NTA.

For each neighborhood, the analysis identifies the category whose local share most exceeds its citywide share. It does this at two levels of detail. A broad category might be “Sanitation and Dumping,” “Parking, Driveways, and Vehicles,” or “Buildings and Construction.” At the more specific level, labels include “Traffic: Drag Racing,” “Appliance: Refrigerator,” and “Lost Property in Taxi or For-Hire Vehicle: Wallet.” The map uses color for the leading broad category and labels each neighborhood with its most distinctive specific category.

The method gives the map its comic force. Healy describes an “Upper East Side to Lower Manhattan” nightlife corridor where residents report lost wallets or keys, alongside brownstone Brooklyn’s bicycle complaints and South Brooklyn reports involving curb cuts and illegal carports. Woodside, Elmhurst, and Corona stand out for public drinking or abandoned vehicles. Forest Hills is associated with “Other Non-Emergency Police Matters,” a broad category that leaves plenty to the imagination.

A map of reporting as well as place

These labels invite the reader to “Tag Yourself” as a neighborhood stereotype. That’s part of the appeal, but the map needs to be read with care. It shows differences in the composition of complaints, not a complete account of neighborhood life. A category can stand out because it occupies a larger share of reports locally than it does citywide; that doesn’t mean it is the area’s most common problem in absolute terms.

The records also reflect who reports, how easily they can report, and how the city defines and records categories. The analysis limits the period, harmonizes categories, and counts addresses to address some of those challenges. Those choices make the comparison more interpretable, but they can’t turn service requests into a direct measure of every underlying condition.

That distinction is what makes the project more than a collection of neighborhood jokes. The map is a compact demonstration of how a different denominator can reveal patterns that raw totals obscure—and of how administrative data carries traces of both the issue being reported and the process used to report it. Read it as a portrait of what residents bring to 311, and the city’s local quirks become visible without mistaking the reports for the whole story.