Methodology

Source

Everything comes from the New York Times Article Search API, which indexes Times articles from 1851 on. Data provided by The New York Times. The pull ran on Sept. 23, 2026 (Park Slope, Williamsburg, Harlem, Bushwick and 10 more everyday names were added the same evening), and covers articles through Sept. 22, 2026.

The search

Each name was searched as a quoted phrase together with the name of its borough, for example "Dumbo" Brooklyn. The API treats the two terms as both required. Adding the borough cuts out most unrelated uses (a Disney elephant, London's Soho, an Adirondack Cobble Hill), but not all of them.

For each name the API returned the hit count for nine periods (before 1950, then each decade from the 1950s to the 2020s) and the 20 oldest matching articles. The 2020s run only through Sept. 22, 2026.

What the counts mean

• They are raw article counts. The Times published a different number of articles in each decade and its archive coverage varies, so a rising count partly reflects the paper, not just the name.
• The API's matching is looser than an exact phrase match. Some older articles match because of scanning errors, and one 1970 article about SoHo matched a search for SoHa.
• A match means the words appear somewhere in the article, not that the article is about the neighborhood. Williamsburg's counts also include articles about East Williamsburg and the Williamsburg Bridge.
• The API reports at most 10,000 matches for any one search. The pre-1950 counts for Harlem, Long Island City, Flatbush, Coney Island, Flushing and Jamaica hit that ceiling, so they are shown as 10,000+ and are floors. Astoria's counts likely include some articles about the Waldorf-Astoria hotel.

First use

For each name, the oldest matching articles were read by hand, using the headline and the archive's summary. The first use is the oldest article whose headline or summary shows the name used for the place. Most older articles have no summary text, so the first use is often later than the oldest match. Where no headline or summary showed the name, the page uses the earliest match that wasn't clearly something else and marks it with an asterisk. Date-bounded follow-up searches were run for SoHo (from 1960 and 1969), Two Bridges, Manhattan Valley (from 1950), Dumbo (from 1988) and East Williamsburg (from 1960) where the oldest 20 matches were all other uses.

On the map and in the "most used" list, matches from decades before a name's first use are set to zero, because they are other meanings. The chart shows them in gray.

Groups

The groups follow fixed rules:

• In the Times before 1950. The name's pre-1950 matches were accepted as this place, either because a pre-1950 headline confirms it (Hell's Kitchen, Harlem, Williamsburg, Bushwick, Park Slope, Prospect Heights, Manhattan Valley, Stuyvesant Heights, Upper East Side, Lower East Side, Brooklyn Heights, Bedford-Stuyvesant, Red Hook, Gowanus, Astoria, Long Island City, Crown Heights, Flatbush, Bay Ridge, Bensonhurst, Canarsie, Sheepshead Bay, Brighton Beach, Jackson Heights, Forest Hills, Flushing, Rego Park, Far Rockaway, Riverdale, Mott Haven, Hunts Point, City Island, Tottenville) or because there were at least 20 pre-1950 matches for a specific place name that has no common other meaning (Hunters Point, Port Morris, Dutch Kills, Ditmas Park, Upper West Side, Greenwich Village, Greenpoint, Kensington, Dyker Heights, Brownsville, East New York, Coney Island, Sunnyside, Ridgewood, Jamaica, St. George, Great Kills). For Kensington and Brownsville some early matches are likely London's Kensington and Brownsville, Texas. These names were then split: faded, then came back if the quietest decade of the 1950s through the 1970s had no more than 20 articles and no more than 5% of the name's busiest decade since 1950, and in steady use otherwise. The 20-article floor was added when Park Slope was included: its 124 articles in the 1950s were under 5% of its 2000s peak, but it had not faded.
• Never caught on. No pre-1950 use accepted, and no decade since 1950 reached 50 articles.
• Took off after 1950. Everything else.

Chelsea was considered and left out: the search cannot separate it from London's Chelsea or from Chelsea Clinton. Pre-1950 matches for Sunset Park (other Sunset Parks and the park itself), SoHo, Cobble Hill, Hudson Square, Greenwood Heights and East Williamsburg were not accepted. The first two are mostly other places. For the other three, the matches could not be checked.

Names left out

Eighty-one names were searched; 79 are shown. Two were dropped because the search could not separate them from ordinary words: NoMad (the word "nomad" swamps it in every decade) and Two Bridges (the phrase turns up in bridge and transit stories).

Map

The map covers all five boroughs and can be zoomed. At the widest view, names that cannot be labeled without overlapping show as dots until you zoom in. Label positions are hand-placed approximate centers, not official boundaries. The base map uses New York City borough, 2020 neighborhood tabulation area and parks boundaries from NYC Open Data, clipped to the shoreline, and U.S. Census county and water files for New Jersey.

Reproducing it

The scripts are in the repository: build/fetch.py runs the searches, build/followup.py runs the date-bounded searches, build/review.py holds every hand decision and the group rules, and build/assemble.py writes data/names.json. Raw API responses are kept locally, not published, in line with the Times API terms. Anyone with a free Times API key can re-run the pull.

Every name