Map any part of India at 10 m, in the classes you need.
Start from a base map of greenery, water, built-up and barren land. Split any class into finer ones, using a model someone already trained or a few polygons of your own, and the map follows. Every model is shared in a model zoo so the next person starts further ahead.
A short walk through: load a base map, split greenery into acacia with a zoo model, run it over Sanjay Van, save the output.
How it works
Start a project
Pick an area and a year. Choose the base classes to start from: IndiaSAT's four or WorldCover's seven.
Split the classes you care about
Use a model from the zoo and say what its classes mean to you, or upload labelled polygons
(one GeoJSON, a class on each) and train your own in minutes.
Run, keep, share
Every run is saved with the scheme that made it. Download the GeoTIFF or the whole project, and make it public when it's ready.
Under the hood: Google's Alpha Earth embeddings, classified as band math inside Earth Engine, so a whole district renders as map tiles with nothing downloaded.
Public projects
Loading…
CoRE stack: Commoning for Resilience and Equality
The CoRE Stack is an open initiative dedicated to empowering communities and researchers with geospatial data tools for sustainable land and natural resource management. Centered on collaboration and participation, it is intended to contribute towards scientific decision-making, climate resilience, and improved social-ecological sustainability. A large network of computer scientists, open-source product developers, environmental research organizations, civil society organizations, and philanthropic donors have come together to build the CoRE stack as a truly digital public infrastructure built and managed by the public.