Closed romulogoncalves closed 6 years ago
We should plot Kmeans results using the centroid's value instead of the ID. Then we should sort on centroid value and using the order to pick the colors.
Like this it is easier to compare maps to see if spatially close clusters are related.
Covered by the new versions in the pull request: https://github.com/phenology/infrastructure/pull/56
We should plot Kmeans results using the centroid's value instead of the ID. Then we should sort on centroid value and using the order to pick the colors.
Like this it is easier to compare maps to see if spatially close clusters are related.