NHGIS Download and Formats: Historical U.S. Census Data, Boundaries, and GIS Workflows

What NHGIS contains

NHGIS organizes data by source, year, topic, geographic level, and table type. Its availability documentation lists decennial census summary tables from 1790–2020, ACS 5-year periods from 2005–2009 through 2020–2024, ACS 1-year periods from 2010 through 2024 except 2020, historical County Business Patterns, vital statistics, agricultural and religious censuses, manufacturing census data, environmental summaries, annual tract population estimates, time-series tables, and geographic crosswalks. Coverage varies by source and geography; a topic available for counties in one period may not exist for tracts or places in another. (nhgis.org)

paragraphs_type_note

• •Summary tables: source-specific statistical tables for census, ACS, and selected historical programs.
• •Time-series tables: selected measures linked across multiple years and geographic levels, with documentation describing comparability and construction.
• •GIS files: mostly polygon boundaries for census reporting areas, plus centers of population and place points.
• •Geographic crosswalks: files describing correspondence between census geographic units from different years.
• •Supplemental resources: environmental summaries and annual tract estimates delivered outside the standard Data Finder workflow.

Geography, temporal coverage, and update cadence

NHGIS is geographically limited to the United States and associated U.S. census geographies represented in its sources. Available units include nation, regions, divisions, states, counties, county subdivisions, places, census tracts, block groups, blocks for selected periods, metropolitan areas, Public Use Microdata Areas, school attendance areas, watersheds, and other source-specific levels. The exact availability and coverage vary by year. For example, historical census tracts did not cover the entire country in the earliest periods, and some specialized boundary products cover selected areas only. (nhgis.org)

paragraphs_type_note

• •Temporal span: decennial census tables extend from 1790 through 2020; ACS availability listed by NHGIS extends through 2024.
• •Boundary interpretation: the year associated with an NHGIS GIS file generally identifies the survey or tabulation year, not necessarily the year in which the boundary geometry was created.
• •Boundary basis: NHGIS states that 2000 and earlier boundaries are generally based on 2000 TIGER/Line files, while 2010 and later boundaries generally use the TIGER/Line release of the same year, subject to documented exceptions.
• •Update cadence: NHGIS does not publish a simple fixed global refresh schedule. Releases occur as source datasets, corrections, boundary files, time-series tables, and project revisions are incorporated. The revision history records version changes and corrections; the current citation page identifies Version 21.0, published in 2026. (nhgis.org)

Download options, formats, and access methods

The NHGIS Data Finder is the primary access point. Users can browse available data without logging in, select datasets and tables by year, topic, geographic level, and source, add GIS boundary files, choose an output layout, and submit an extract request. A free registered account is required to submit the request. NHGIS processes the request on its server and makes the resulting files available through Extracts History, also notifying the user when the extract is ready. Extracts are available for download for a limited period, after which users can regenerate them. (nhgis.org)

paragraphs_type_note

• •CSV: comma-delimited tabular output is the practical choice for GIS joins, spreadsheets, and general data processing. Field names may use compact NHGIS codes, so retain the accompanying codebook.
• •Fixed-width: suitable for statistical packages and accompanied by command files for software such as SPSS, Stata, and SAS.
• •Shapefile: NHGIS supplies GIS boundary and point data in the shapefile format, normally inside ZIP archives. Keep all component files together when extracting or moving a shapefile.
• •Documentation: extracts include codebooks, readme files, table descriptions, and citation information needed to interpret fields and values.
• •Programmatic access: IPUMS documents an API, and the NHGIS user guide identifies the ipumsr and ipumspy packages for R and Python workflows. Supplemental resources such as environmental summaries and annual tract estimates can also be accessed through the IPUMS API. (nhgis.org)

Suitable GIS and spatial-analysis workflows

A reliable NHGIS workflow begins with the geography, not the map. First define the unit of analysis—such as county, tract, place, or metropolitan area—and the year or period to be compared. Then select a table whose universe, concepts, denominators, and statistical definitions match the question. Download the associated GIS file for the same tabulation geography and inspect its year, source, identifiers, projection information, and boundary notes before joining data. NHGIS documentation emphasizes the role of GISJOIN identifiers in linking tabular and spatial files, while recent Census-derived files may also contain GEOID-related identifiers. (nhgis.org)

paragraphs_type_note

• •Choropleth mapping: join one or more NHGIS measures to tract, county, place, or state polygons, then calculate rates or percentages only after confirming the correct denominator.
• •Historical comparison: use NHGIS time-series tables or geographic crosswalks where appropriate, but distinguish genuine temporal change from changes in geographic boundaries, definitions, survey design, and source tables.
• •Feature enrichment: spatially join buildings, roads, transit stops, service locations, environmental observations, or infrastructure to NHGIS areas, then summarize counts, rates, or coverage by geography.
• •Reproducible extraction: record the NHGIS version, extract date, source dataset, table identifiers, geography year, download format, and codebook alongside the analysis.
• •Large-table processing: prefer compact CSV output for GIS joins, use codebooks to decode fields, and consider R, Python, or database-based processing when tables contain many variables or geographic levels.

License, strengths, and tradeoffs

NHGIS grants users a limited license to use its data and documentation. The published terms prohibit redistribution without permission and require appropriate citation. The project provides a recommended citation for Version 21.0 and asks users to add publications and educational materials to the IPUMS bibliography. Analysts should review the current terms before embedding NHGIS data in a public product, redistributing extracts, or using the data commercially. (nhgis.org)

• •Strength: unusually broad historical depth, including census data reaching back to 1790 and boundary resources for many changing U.S. geographic units.
• •Strength: a single discovery and extraction workflow combines statistical tables, time-series resources, crosswalks, documentation, and GIS files.
• •Strength: formats and identifiers are practical for joining tabular data to boundaries and then connecting area-level statistics to other spatial datasets.
• •Tradeoff: NHGIS is area-based statistical data. It generally cannot answer questions about individual addresses, parcels, buildings, roads, or households without another dataset and an appropriate geographic linkage.
• •Tradeoff: boundary vintages, changing definitions, historical coverage gaps, and source-specific universes can make naïve year-to-year comparisons misleading.
• •Tradeoff: redistribution is restricted under the published terms, and the default shapefile format may require conversion for modern cloud-native or database workflows.
• •Tradeoff: ACS and other sample-based estimates include uncertainty; users should preserve margins of error where available and avoid presenting estimates as exact counts.