Two joinable datasets making every US place Wikidata knows analytics-ready: 81,580 places (cities, towns, villages, boroughs, townships, CDPs and neighborhoods, rostered by the US-only FIPS identifier so nothing needs a class list) and a 112,516-row historical census panel behind them, 1790 to 2020. The gazetteer exists for the JOINS: ZIP codes on 28,556 places, FIPS on 52,434, GNIS identifiers, so an analyst can attach any ZIP- or FIPS-keyed data (home prices, health, weather) to a clean spine. Around the keys: state and county resolved from the administrative chain, coordinates, elevation, area (taken as the LARGEST stated figure, because cities annex outward and the occasional bogus small statement must lose), derived density, founded year, what the place is named after, and the population story: latest count, the 2010 and 2020 census points with the growth percent between them (computed for 48,291 places), and each place's census-series depth. The history file is the panel: Detroit's 1950 peak and long fall, every Sun Belt boomtown's curve, one row per place per census on the census-beats-estimate rule. Wikidata holds no census socioeconomics; income and demographics live at the Census Bureau, and the ZIP and FIPS columns exist precisely so you can join them in. Built with a 16-test acceptance suite anchored on New York's crown, Chicago in Cook County, Detroit's peak and exact growth arithmetic.
194,096 rows across 2 joinable files, with 30 columns in total: Place, Place_ID, Type, State, County, ZIP_Codes, FIPS, GNIS_ID, Latitude, Longitude, Elevation_M, Area_KM2, Founded_Year, Named_After, Latest_Population, Latest_Pop_Year, Pop_2020, Pop_2010, Growth_2010_2020_Percent, Density_Per_KM2, Census_Observations, Earliest_Census_Year, Place_URL, Place, Place_ID, State, Census_Date, Year, Population, Method.
It is built by programmatically scraping wikidata.org, last pulled on 2026-09-14. Datasets are versioned; older versions stay downloadable.
Yes. Every CodeSights dataset is completely free as a CSV download; a free account is all it takes. Anyone can preview the data without signing in.
Yes. The exact Python scraper that built it is viewable on the dataset page by any signed-in member, so every number is reproducible.
Join the files on Place_ID (history.Place_ID = places.Place_ID); Census_Observations on the places file equals that place's history rows exactly. ZIP_Codes and FIPS are the outward join keys for your own data.
Automated scraping leaves room for error and the underlying sources change over time, so no version is guaranteed accurate or complete. If a number matters, verify it against the original source.