Just checked and that repo had much less data in 2019 than the other years, hence the jump. Repo size/volume of activity is indeed something that varies a lot between these repos - but I didn't control for the intentionally because I wanted to see results not just for big, well-funded OSS, but also community/volunteer-driven ones. I think it backs up the argument that we need to be doing more to fund OSS in general.
Would be great to see run this analysis on some other OSS projects! Are you going to be open sourcing the code?