NHGIS: Historical U.S. Census Data, Boundaries, and GIS Analysis Workflows
IPUMS NHGIS, the National Historical Geographic Information System, provides downloadable U.S. census and survey summary tables, historical time series, geographic crosswalks, and GIS-compatible boundary files. Its coverage extends from the 1790 Census through recent American Community Survey releases, with data for states, counties, places, census tracts, block groups, blocks, school districts, legislative districts, metropolitan areas, and other geographic units. NHGIS is especially valuable for studying demographic and socioeconomic change over time, joining historical statistics to geographic boundaries, and building reproducible U.S.-focused spatial-analysis workflows. It is not a global basemap, address database, current parcel layer, road network, or continuously updated inventory of physical features.
NHGIS is a data access and harmonization system for U.S. geographic census information. It brings together summary tables, time series, historical and contemporary boundaries, and crosswalks so analysts can connect population, housing, economic, agricultural, environmental, and related statistics to geographic units across time. The project is maintained by IPUMS at the University of Minnesota and provides access through a custom extract system rather than a single static download.
How EDK fits
For: Researchers, public agencies, policy analysts, historians, demographers, GIS professionals, journalists, students, and developers building U.S. location-data or spatial-analysis workflows.
NHGIS connects to E2’s physical-world perspective at the level of geographic context rather than feature discovery. NHGIS describes people, households, housing, employment, agriculture, and other measured characteristics within named U.S. areas; E2’s world-model perspective is concerned with finding and computing about the places, infrastructure, environmental features, and public assets that occupy those areas. In a combined workflow, NHGIS can provide the population or socioeconomic context for a region, while a geographic computation layer can relate those statistics to roads, buildings, transit, waterways, public facilities, service requests, or other physical features. The important distinction is that NHGIS supplies statistical attributes and historical geographic frameworks; it does not itself provide a universal representation of every physical object on Earth.
- •Join NHGIS population, housing, or income statistics to geographic regions and compare them with nearby buildings, roads, public facilities, environmental features, or infrastructure when evaluating spatial access, exposure, service coverage, or neighborhood change.
- •Use NHGIS historical boundaries and time series as the demographic context for geographic computations such as ranking areas, measuring proximity to physical assets, aggregating features by census geography, or analyzing how the surrounding population of a place has changed over time.
- •Combine NHGIS data with current physical-world layers to distinguish statistical change from physical change—for example, comparing shifts in population or housing characteristics with new construction, road access, vegetation, waterways, or public-service activity.
Dataset profile
Formats
Esri Shapefile boundary and point files with associated metadata
Comma-delimited tabular extracts
Fixed-width tabular extracts for selected products
NHGIS-specific GISJOIN and geographic identifiers
API responses and programmatic access through IPUMS developer resources
R and Python workflows through documented IPUMS packages
Access methods
NHGIS Data Finder custom extract system
NHGIS geographic crosswalk download pages
Direct pages for selected environmental summaries, annual tract estimates, and special boundary products
NHGIS API resources documented through the IPUMS Developer Portal
R package ipumsr
Python package ipumspy
Social Explorer online mapping for some NHGIS content
Quality and limits
Not a global dataset; geographic coverage is U.S.-focused.
Not an address, parcel, road, building, infrastructure, or continuously updated physical-feature inventory.
Historical boundaries may be reconstructed from older TIGER/Line files, printed maps, digitization, or other sources.
Geographic units, codes, definitions, and boundaries change over time.
ACS values are sample-based estimates and should be interpreted with margins of error and survey-period limitations.
Some historical geographic units have incomplete spatial coverage, especially in early census years.
The project’s license restricts redistribution without permission.
A table’s universe, classification, source, and geographic integration method must be checked before comparison or aggregation.
Analysis workflows
Historical demographic mapping by county, tract, place, or other census geography.
ACS housing, income, population, employment, or social-indicator analysis joined to boundary files.
Time-series analysis using nominally integrated or geographically standardized tables.
Spatial joins between NHGIS statistics and roads, buildings, transit, public facilities, watersheds, environmental layers, or infrastructure.
Regional aggregation and ranking by census geography.
Population or socioeconomic context around physical features, service areas, or infrastructure corridors.
Crosswalk-based translation of statistics between geographic vintages.
Reproducible extraction and analysis using the Data Finder, API, R, Python, or GIS software.
Value inside EDK
What NHGIS contains
NHGIS is broader than a single census download. Its summary tables include decennial Census data from 1790 through 2020, American Community Survey estimates, County Business Patterns, vital statistics, agricultural census data, religious bodies data, manufacturing census data, and selected environmental and watershed summaries. The project also provides time series tables that link comparable statistics across multiple census years. (nhgis.org)
The subject matter is primarily demographic, social, housing, economic, agricultural, and historical. Depending on the source and year, tables may describe total population, age, sex, race, ethnicity, households, housing occupancy, tenure, education, income, poverty, employment, business activity, agricultural characteristics, births, deaths, or other measured attributes. NHGIS is therefore best understood as a statistical and geographic reference system, not as a conventional physical-feature dataset.
- •Decennial Census summary data from 1790–2020.
- •ACS 1-year and 5-year summary data through recent releases, including 2024 ACS products and 2020–2024 ACS 5-year data.
- •Time series tables with nominally integrated and geographically standardized options.
- •Historical and contemporary GIS boundaries, point files, and geographic crosswalks.
- •Selected non-Census sources such as County Business Patterns, vital statistics, agriculture, religious bodies, and environmental summaries.
Geography, time coverage, and update cadence
NHGIS is limited geographically to the United States and associated U.S. census geographies; it is not a global demographic database. Its historical span is unusually long: some state and county population tables reach back to 1790, while many other subjects begin later. Decennial Census coverage extends from 1790 through 2020, and ACS coverage includes 5-year periods from 2005–2009 through 2020–2024 and 1-year periods from 2010 through 2024, excluding the standard 2020 ACS 1-year release. (nhgis.org)
There is no single universal NHGIS update interval. Updates depend on the source program and release cycle. ACS products are added as the Census Bureau releases new summary files; decennial data arrive on the decennial schedule; and boundary files, documentation, crosswalks, and special products may be updated separately. NHGIS publishes substantial changes through its revision history. As of the current documentation, the site identifies 2024 ACS 1-year data and 2020–2024 ACS 5-year data among its recent additions. (nhgis.org)
- •Long historical coverage, but not every topic or geography exists for every year.
- •ACS estimates represent survey periods and estimates, not complete administrative counts.
- •Boundary years may identify the survey geography rather than the exact source-file vintage.
- •Update timing varies by source dataset and NHGIS product.
Formats and access methods
NHGIS uses a custom extract workflow. Users browse the Data Finder, select summary tables, time series tables, geographic levels, years, geographic extents, and optional GIS boundary files, then submit an extract request. NHGIS generates the requested files on its servers and sends an email when the extract is ready. A free account is required to submit requests and download extracts, although users can explore available data before logging in. (nhgis.org)
GIS files are primarily distributed as Esri Shapefiles, including the associated component files and XML metadata. Most represent polygon boundaries, while some products provide points such as centers of population or functional place centers. NHGIS commonly adds a GISJOIN attribute so boundary files can be joined to NHGIS tabular data. The project’s boundary files are generally projected into the USA Contiguous Albers Equal Area Conic coordinate system, although users should verify the metadata for each extract. (nhgis.org)
- •Data Finder custom extracts rather than one universal file.
- •Shapefile boundary and point data with NHGIS-specific identifiers and metadata.
- •Tabular outputs for spreadsheet, statistical, database, and GIS workflows.
- •API, R, and Python access options documented through IPUMS resources.
- •Online mapping is available through Social Explorer for some NHGIS content, but it is separate from local GIS downloads.
Suitable GIS and spatial-analysis workflows
A defensible NHGIS workflow begins by defining the geographic question. Decide whether the analysis needs a current ACS estimate, a decennial count, a historical census table, a time series, or a standardized geography. Then identify the correct universe, classifications, geographic level, year, and source dataset. A population table can mean all residents, a specific age group, a race or ethnicity category, households, housing units, or another defined universe; those distinctions must remain visible in the analysis.
For mapping, request the matching GIS boundary files and join the table using the supplied GISJOIN or documented geographic identifier. Validate row counts, check for unmatched records, inspect missing or special values, and confirm that the boundary year corresponds to the tabular data’s geographic definition. For longitudinal analysis, use the relevant time series product where available and preserve the documentation describing integration or standardization methods.
- •Select tables by universe, topic, source, year, and geographic level before downloading.
- •Use matching boundary definitions and documented identifiers for joins.
- •Apply margins of error and survey interpretation rules to ACS estimates where relevant.
- •Use crosswalks or standardized time series when comparing changing geographies.
- •Combine NHGIS with current physical-world layers for feature-level or infrastructure analysis.
License, strengths, and tradeoffs
NHGIS data and documentation are available free of charge, but the project’s terms are not an unrestricted public-domain dedication. The published license requires appropriate citation and prohibits redistribution without permission, while allowing users to request permission for particular redistribution or commercial uses. Projects that package NHGIS extracts into public products should review the current terms rather than assuming that a downloaded file can be republished freely. (nhgis.org)
Boundary construction also varies by period. Recent files are mainly based on Census Bureau TIGER/Line files with NHGIS modifications, while older boundaries may involve historical maps, other TIGER/Line vintages, manual digitization, and additional editing. This makes NHGIS valuable for historical research, but it also means that a boundary should not automatically be interpreted as a modern legal, cadastral, or precisely surveyed polygon. (nhgis.org)
- •Strength: exceptional historical depth for U.S. population and geographic analysis.
- •Strength: combines statistics, time series, boundaries, and crosswalks in one documented workflow.
- •Strength: supports both GIS users and programmatic analysis through extracts, API resources, R, and Python.
- •Tradeoff: coverage, definitions, and geographic detail vary by source and year.
- •Tradeoff: historical boundary construction can involve approximation, digitization, or non-equivalent vintages.
- •Tradeoff: license terms restrict redistribution, and ACS estimates require statistical caution.
Questions teams ask
Is NHGIS the same as the U.S. Census Bureau data portal?
No. NHGIS is an IPUMS project at the University of Minnesota that assembles, documents, harmonizes, and delivers U.S. census and related geographic data from sources including the Census Bureau. It provides its own table organization, time series, boundary files, identifiers, extract workflow, and documentation.
Can NHGIS provide data for a single city, county, or small area?
It depends on the product. NHGIS generally delivers files for a geographic level across the United States or selected states rather than arbitrary extracts limited to one county or city. Some table data can be requested for selected states, and block or block-group GIS files can be obtained for selected states. Users should check the Data Finder and FAQ for the specific geography.
Are NHGIS boundaries suitable for current parcel or address analysis?
No. NHGIS boundaries represent census, survey, administrative, or historical geographic units. They are useful for aggregation and spatial context but are not parcel boundaries, address points, building footprints, or a guarantee of current legal geography. Historical files may also be reconstructed from maps and other sources.
How should I compare NHGIS statistics across years?
First confirm that the table measures the same concept, universe, classification, and source type in each year. Then determine whether the series is nominally integrated or geographically standardized, inspect boundary changes, and preserve the relevant documentation. For ACS estimates, also account for the survey period, margins of error, and changes in published variables.
Can I redistribute an NHGIS extract?
The published NHGIS terms state that users may not redistribute data without permission and must cite NHGIS appropriately. A subset may be published to meet journal-access requirements, while other redistribution or commercial uses should be discussed with NHGIS. Review the current citation and use terms before republishing files.
Keep exploring
census acs
Census ACS is a closely related source for current U.S. demographic, housing, social, and economic estimates that can be used alongside NHGIS historical and geographic data.
gadm
GADM provides a contrasting global administrative-boundary dataset useful for understanding how NHGIS’s U.S.-specific census geography differs from worldwide boundary products.
Sources
IPUMS NHGIS home page
Official overview of NHGIS coverage, historical span, subject areas, and access to U.S. census data and mapping files.
Open sourceNHGIS Data Availability
Official availability tables for summary data, time series, crosswalks, GIS files, geographic levels, and years.
Open sourceNHGIS GIS Files
Official documentation for Shapefile formats, geometry years, GISJOIN identifiers, boundary derivation, and historical limitations.
Open sourceNHGIS Citation and Use of NHGIS Data
Official license, citation, and redistribution terms.
Open sourceNHGIS Data Documentation
Official documentation for tabular sources, time series construction, GIS files, and revision history.
Open source