This is a super simple example of clustering a set of polygons with attributes into a defined number of classes using KMeans.
The notebook reads the spatial data (within a bounding box), allows selection of clustering parameters, and computes the silhouette score to help determine the optimal number of classes.
The resulting cluster number can be saved into a new spatial file.
To check out the data you'll need DVC installed with Cibo Labs public S3 remote added:
dvc remote add -d cibo-dvc s3://cibo-dvc/
Or grab the data from OneDrive.
Example output:
