Where to Get Free Geospatial Data: 10 Sources Every GIS User Should Know

Why this list exists

Free geospatial data is everywhere on the internet, but the quality and longevity of sources varies wildly. This list focuses on the ten that have proven durable: well-maintained, openly licensed, and used across enough projects that you'll meet their formats again and again.

For each one: what's in it, how to download, format, typical use case.

1. OpenStreetMap (planet + Geofabrik regional extracts)

The world's largest open vector geodatabase: roads, buildings, points of interest, land use, addressing, the lot.

• Download: planet.osm.org for the full planet (~75 GB compressed); Geofabrik for country and regional extracts (much smaller).
• Format: .osm.pbf (Protocol Buffers binary). Also available as .osm XML.
• Conversion: osmium tags-filter and osmium extract for slicing, then ogr2ogr for format conversion. Direct PBF→GeoJSON is a route for small extracts only.
• License: Open Database License (ODbL).
• Use case: any project needing detailed vector data for a specific region — basemaps, network analysis, address geocoding.

2. Natural Earth

Cleaned, generalised vector and raster data at three scales (1:10m, 1:50m, 1:110m). Country boundaries, populated places, rivers, roads, cultural features.

• Download: naturalearthdata.com.
• Format: Shapefile (vector), GeoTIFF (raster).
• License: Public domain.
• Use case: small-scale maps (continental or global), reference layers, schematic cartography.

3. US Census TIGER/Line

Authoritative US administrative and statistical boundaries: states, counties, census tracts, blocks, congressional districts, roads, hydrography.

• Download: census.gov/geographies/mapping-files.
• Format: Shapefile per geography, per year. Annual vintages back to ~2007.
• License: Public domain (US federal work).
• Use case: any US-based spatial analysis tied to census data; demographic mapping; redistricting work.

4. Copernicus (Sentinel, Land Monitoring)

European Space Agency open-access programme. Sentinel-1 SAR, Sentinel-2 optical (10 m resolution), Sentinel-3 ocean and land, plus the Copernicus Land Monitoring vector products (Urban Atlas, CORINE Land Cover).

• Download: dataspace.copernicus.eu for raw imagery; land.copernicus.eu for vector products.
• Format: SAFE / JP2 for imagery, GeoPackage / Shapefile for vector products.
• License: Free for any use with attribution.
• Use case: environmental monitoring, agricultural analysis, land use change detection, urban morphology in Europe.

5. EU Open Data Portal

Aggregator for EU institutions and member-state public data. Lots of spatial layers across transport, agriculture, environment, statistics.

• Download: data.europa.eu.
• Format: mixed — Shapefile, GeoJSON, GeoPackage, CSV, sometimes WMS / WFS endpoints.
• License: mostly CC-BY or equivalent.
• Use case: EU-wide policy mapping, cross-country comparative analysis, regulatory geography.

6. OpenAerialMap

Open aerial and drone imagery, contributed by humanitarian mappers, governments, and individuals. Useful for high-resolution basemaps in areas Sentinel-2 doesn't reach.

• Download: openaerialmap.org.
• Format: GeoTIFF (COG when available), tile URLs.
• License: CC-BY 4.0 or compatible.
• Use case: post-disaster mapping, agricultural inspection, situations needing sub-metre imagery.

7. SRTM DEM

Shuttle Radar Topography Mission digital elevation model. Near-global elevation data at 1 arcsec (~30 m) resolution.

• Download: USGS Earth Explorer (registration required) or via Copernicus.
• Format: GeoTIFF tiles, HGT files.
• License: Public domain.
• Use case: terrain analysis, watershed modelling, viewshed calculation, hillshade basemaps.

8. GADM

Global Administrative Areas — country, state/province, district, and sub-district boundaries for every country, at multiple administrative levels.

• Download: gadm.org.
• Format: GeoPackage, Shapefile, GeoJSON, KMZ.
• License: Free for academic and non-commercial use; commercial use requires permission.
• Use case: consistent multi-country administrative mapping where harmonised boundaries matter more than authoritative national data.

9. Overpass API

A read-only query API over OpenStreetMap data. Lets you pull specific tag/region combinations without downloading the full PBF.

• Endpoint: overpass-api.de/api/interpreter, or the friendly web frontend Overpass Turbo.
• Format: results as JSON, XML, CSV, KML, or GeoJSON.
• License: ODbL (inherits from OSM).
• Use case: quick targeted extractions ("all cafes in Berlin," "all level crossings in Germany") without setting up a full OSM pipeline.

10. World Bank Spatial Data

Tabular and spatial datasets from World Bank development indicators, often joined to country or sub-national boundaries.

• Download: datacatalog.worldbank.org (filter by Type: Geospatial).
• Format: Shapefile, GeoJSON, CSV with country codes.
• License: CC-BY 4.0.
• Use case: development geography, cross-country socioeconomic mapping, infrastructure indicators.

Honourable mentions

• GBIF for species occurrence data — global biodiversity records as CSV with lat/lon.
• EU-DEM for European-specific elevation (better than SRTM in Europe).
• OpenStreetMap-derived basemap providers (Stadia Maps, MapTiler, Maptiler) for hosted tiles when you don't want to render your own.
• City-level open data portals (Berlin's daten.berlin.de, NYC's NYC Open Data) for fine-grained urban data.

A note on licensing

Reading the license matters more than the format. OpenStreetMap's ODbL has share-alike terms — derivative datasets must also be ODbL-licensed if published. Natural Earth and US Census TIGER are public domain. Copernicus is permissive but requires attribution. GADM restricts commercial use.

When in doubt, attribute the source, link the license, and ask the publisher if your use case is novel.

Recommended starting kit

For a generic GIS project getting off the ground:

1. OpenStreetMap (via Geofabrik extract) for detailed regional vector.
2. Natural Earth for global reference.
3. SRTM for terrain.
4. GADM or local authoritative source for admin boundaries.

That's four data sources covering 80% of typical GIS authoring needs. Layer in the others as the project demands.