Normalize package data to consider the packages that have LARGE numbers compared to most of our packages that are smaller Clean up pyos package metrics

[lwasser]

We current have this page that attempts to show the "size" and activity of packages in our ecosystem

https://www.pyopensci.org/metrics/peer-review/pyos-package-dashboard.html

One of the challenges of these plots is that packages like astropy are so large it overpowers the overall metrics.
If you have a eye for analytics and plots, it would be good to think about how we can present this data in a way that astropy and our few bigger packages don't over weight the general numbers that represent our packages

See the image below

the data for the packages come from this file:
https://github.com/pyOpenSci/pyopensci.github.io/blob/main/_data/packages.yml