Points of interest layers question

[cornhundred]

Hi I'm new to tippecanoe and I have a question about using layers for point data with different categories. I have a CSV file with millions of points of interest that each belong to one of a few hundred different categories. I would like to use the --cluster-distance option to merge points into clusters, but I would only like to merge points from the same category. Can this be done by having points from each category sent to their own layer? And if so, do I need to split my points into different files for each category and build a mbtiles file where multiple sources represent multiple layers or can I point to the category column in the CSV and get tippecanoe to know that these points should be put into different layers?

Alternatively, would it be preferable to cluster points from all categories all points and keep track of the breakdown of each category type in each cluster (e.g. using --accumulate-attribute for hundreds of attributes/categories)?