A remarkable share of the world’s geospatial foundation is free: Landsat and Sentinel imagery archives, near-global elevation models, OpenStreetMap’s crowd-built vector map, administrative boundaries, climate reanalyses and a growing family of AI-derived layers such as global building footprints. Knowing these sources — and their licences and limitations — is basic literacy; entire products and national programmes stand on them.
The imagery and elevation pillars
Earth observation: Landsat (five decades, 30 m) and Sentinel-1/2 (radar and 10 m optical, days-scale revisit) anchor free imagery, accessible through open portals and cloud platforms. Elevation: SRTM and Copernicus GLO-30 provide ~30 m global terrain; national LiDAR programmes (like the US 3DEP) publish far finer where they exist.
Each carries fitness bounds — resolution, acquisition dates, known artefacts — that responsible use respects and documents.
Vectors, boundaries and derived layers
OpenStreetMap is the planet’s most complete open vector map — roads, buildings, POIs — with quality that varies by community activity and an ODbL licence whose share-alike terms products must honour. Administrative boundaries (GADM, national portals), census and statistical layers, and open address efforts round out the vectors.
A new generation of derived datasets — global building footprints from satellite AI, land-cover maps like ESA WorldCover, population grids like WorldPop — supplies analysis-ready layers where official data is thin.
Using open data professionally
Open does not mean uniform: professional use validates quality in the study area, records versions and access dates, honours licences (attribution, share-alike), and blends open foundations with commercial or field data where accuracy demands. The economic pattern is consistent — open data for breadth, targeted paid data for depth.
Data-source fluency is itself a service: knowing what exists, what it is good for, and where it will quietly betray you.