Glossary

Plain-language definitions of the terms used across this site. Terms are in alphabetical order, grouped by topic.

Jump to: Open data basics · Getting data with code · Data platforms · Maps and places · File formats · AI and connectors · Government and records · The data directory

Open data basics

Bulk download: Getting a whole dataset as a file, instead of asking for parts of it through an API.
Catalog: A searchable list of datasets on a portal, with a description of each one.
Data site (or data portal): A website where a government or organization publishes its data for anyone to use. NYC Open Data and data.ny.gov are examples.
Dataset: One collection of related data, usually a table. For example, “311 Service Requests” is one dataset.
Dataset ID: The short code that identifies a dataset. On Socrata portals it is two groups of four letters and numbers, like abcd-1234.
Metadata: Information about a dataset rather than the data itself: its title, who publishes it, when it was last updated, and what each column means.
Open data: Data that anyone can use, reuse and share for free, usually published by a government.
Open data portal: See Data site.

API (application programming interface): A way for a program to ask a website for data directly, instead of a person clicking through pages.
API address (or endpoint): The specific web address a program sends requests to.
API key, token or app token: A code you get by signing up, which a program sends with each request. Some APIs require one; others accept one to allow more requests. Keep keys private.
Error 403: A response meaning “forbidden”: the site refused the request. Some sites send it to scripts but not to browsers.
Error 404: A response meaning “not found”: nothing is at that address.
Query: A request for specific data, such as “311 complaints in Brooklyn in 2026.”
Rate limit: A cap on how many requests a site will answer in a given time. Going over it gets requests blocked for a while.
REST: A common style of web API, where each kind of data has its own web address and requests are ordinary web requests.
Undocumented API: An API that a site uses for its own pages but has not published for others. It can work, but it can change or disappear without notice.
User-Agent: A line a program sends with each request saying what it is, such as a web browser. Some servers block requests that don’t look like a browser.

ArcGIS: Esri’s family of mapping software. Many New York counties use it to publish map data.
ArcGIS Enterprise: A version of ArcGIS that an organization runs on its own servers.
ArcGIS Hub: Esri’s tool for building an open data site on top of ArcGIS Online. Many county data sites are Hub sites.
ArcGIS Online: Esri’s hosted mapping service. Organizations publish maps and map layers there.
ArcGIS REST: The API that ArcGIS servers use. Programs can ask it for map layers and the data inside them.
CKAN: Open-source software for running a data catalog. The federal catalog, data.gov, has been built on it.
DCAT-US: A standard way to describe datasets so catalogs can share them. Federal agencies use it, and ArcGIS Hub sites publish a DCAT-US list of their datasets.
Discovery API: Socrata’s API for searching across Socrata catalogs to find datasets.
Esri: The company that makes ArcGIS.
Feature service (FeatureServer, MapServer): An ArcGIS map layer published so that programs can request it, including the data behind each map shape.
Socrata: The software behind NYC Open Data, data.ny.gov and several other portals. It is now part of Tyler Technologies.
SODA (Socrata Open Data API): Socrata’s API for reading data from a dataset.
SoQL (Socrata Query Language): The query language SODA uses to filter, sort and summarize data, similar to SQL.
Tyler Technologies: The company that owns Socrata, now sold as Tyler Data & Insights.

BBL (borough, block and lot): The number New York City uses to identify a tax lot.
BIN (building identification number): The number New York City uses to identify a building.
Community district: One of New York City’s 59 areas, each served by a community board.
FIPS code: A standard federal number for a state or county. New York State is 36.
Geocoding: Turning an address into a location on a map, or into an identifier like a BBL.
GIS (geographic information system): Software and data for making and analyzing maps.
Layer: One set of map data, such as roads or tax parcels, that can be shown on a map.
Parcel: A piece of land with its own tax record.

CSV (comma-separated values): A plain-text table, one row per line. Opens in any spreadsheet program.
GeoJSON: A JSON format for map shapes and their data.
GTFS (General Transit Feed Specification): A standard format for transit schedules, routes and stops. GTFS-Realtime adds live arrival and alert information.
JSON: A text format for structured data that programs read easily.
Shapefile: An older but common file format for map data, made by Esri.
XLSX: A Microsoft Excel spreadsheet file.

AI assistant: A chat tool, such as Claude or ChatGPT, that answers questions in plain language.
Connector (MCP server): A small program that lets an AI assistant use an outside tool or data source, such as a data portal.
Grounding: Making an AI assistant answer from real, checkable sources instead of from memory. See why grounding matters.
Hallucination: When an AI assistant states something false as if it were true, such as a bill number that does not exist.
MCP (Model Context Protocol): An open standard for connecting AI assistants to tools and data. The connectors on this site use it.

311: New York City’s line and app for non-emergency service requests and city information.
Community board: An appointed local body that advises on land use, budget and services for a community district.
FOIL (Freedom of Information Law): New York State’s public-records law. It lets anyone request records from state and local government.
Legistar: The legislative records system that the New York City Council uses for bills, hearings and votes.
Public authority: A state or local public corporation that runs a service, such as the MTA or the Port Authority.

These terms are used in the New York data directory.

Built on: The software behind a data site, such as Socrata or ArcGIS.
How to get the data: Whether a program can pull the data (an API), or whether you download files or search by hand.
Key needed?: Whether you must sign up for a key before using the API.
Live: The site worked when we checked it.
No data site: We found no public data catalog for that place. It may still have a map viewer or a property search.
Not yet confirmed: We could not verify the details. Check before you rely on them.
Offline: The site is gone.
Type of source: Which group a source belongs to: state data sites; state agencies, the Legislature and authorities; counties, cities and New York City agencies; or universities, libraries and nonprofits.

Missing a term, or is a definition unclear? Give us feedback.