3D Land Cover Map Generator

Pick an area on the map, or use the fields below, then generate a land cover export via GitHub Actions.

Generate a Map

Area of interest:

Country:

West (lon) South (lat) East (lon) North (lat)

Upload a .geojson/.json file (optional, fills the box below): GeoJSON (Geometry, Feature, or FeatureCollection):

KML file:

Shapefile, zipped (.shp/.dbf/.shx/.prj bundled in one .zip):

Click a tool above, then click on the map to place points (double-click, or click the starting point again, to finish a polygon; click once for a rectangle's first corner, move the mouse, then click again to finish it).

Output format: Mode: Downsample factor: Map title (optional, PDF only): Email the finished map to (optional):

Only works if the repo owner has configured Gmail sending secrets; otherwise this is silently ignored and you still get the download link below.

Open workflow dispatch on GitHub instead

How it works

Pick an area: draw directly on the map, a country, a bounding box, a pasted/uploaded GeoJSON, or an uploaded KML/shapefile (KML and shapefiles are converted to GeoJSON right here in your browser before being sent).
Click Run workflow now — no account or token needed, the request goes through a small proxy that holds the credentials.
This page polls the run's status live. When it finishes, a link to the run (with the downloadable artifact) appears below the button.

Artifact bytes can't be streamed back into this page for a direct "Save As" — GitHub's artifact download redirects to storage that doesn't allow cross-origin reads, a GitHub-side limitation with no backend-free workaround. The link takes you straight to the run's summary page, where the normal GitHub download button works.

The Python tool uses ESRI Sentinel-2 land cover tiles and Copernicus elevation data, clipped to the selected area.

The hillshade-and-color rendering technique is the same one rayshader (an R package by Tyler Morgan-Wall) uses, reimplemented here in Python.