How to specify colormap for categories when using `ds.count_cat()`

Datashader
1

I’m trying to use a custom mapping from categories to colors when datashading categories with ds.count_cat(). How can I do this, if it’s possible?

I tried using the cmap keyword argument but it seems to be ignored.

Sorry if this has been answered elsewhere, but I couldn’t find the answer after several minutes of checking the docs, GitHub, and StackOverflow.

2

We should probably issue a warning suggesting the right solution. When using count_cat you should use a color_key, either mapping each category to a color explicitly as a dictionary or providing a list of colors at least as long as the number of categories.

3

Thanks! I could submit a PR for that—or perhaps even better, add some documentation—if you could point me in direction of the relevant files?

4

To be clear you’re using datashader directly or via HoloViews?

5

I’m using it via Holoviews (hv.operations.datashader).

6

Along similar lines, am i able to specify colors to polygons from a GeoDataFrame using datashader (rendering 28,000 polygons)?
e.g. agg=df.count_cat(‘category_name’)

I’m currently using agg=ds.mean(‘category_name’) and pointing to numeric classes as a workaround, but this changes dynamically. Perhaps I’m doing this completely backwards? But i love that I can render and zoom in on these polygons! Such beautiful tools you’ve developed!