Why should anyone care how many github stars one library has compared to another?
Github stars are proportional to the size of the community.
You don't want to bet your company on a project with only a few thousand stars, it could die quickly.
I agree there's a decent correlation. There are a few notable exceptions/edge cases. For example, Ruby's rake has just ~1,300 stars:
It'd be an interesting project to compare frameworks/tools by aggregating the stars (or whatever metric) of its major (i.e. filtering out hobby forks and skeleton repos) dependent projects, e.g. the 22K stars of gatsbyjs/gatsby would count towards React's overall reach/popularity.