How many packages in each ecosystem?

[CJ-Wright]

It would be great to have a bar chart showing how many packages are in a certain ecosystem, for instance if a package have a dependency on sklearn, tensorflow, spacy, or pytorch that package is part of the data-science ecosystem, if dependent on numpy or scipy numerical computing.

Assembling these sets of prominent ancestor packages would allow us to know how we are serving each community. We could combine this with the download statistics to understand the evolution of the ecosystems.

Thoughts @conda-forge/core