You can read more about pkgstats on https://pkgstats.archlinux.de/ and mind you, only covers usage numbers in percent, not download numbers. Not everybody will agree to share the information anyway in order to have correct numbers (if that was the case to be implemented at some point).
Then we can think the same for AUR or other repositories - download stats are inaccurate. It does not indicate if the user actually liked and kept the package or not. The votes for AUR packages don’t tell the story of the most popular, because not everyone votes …
In short, from what i understand, what you propose will not happen. Hopefully someone can prove me wrong.