AFAICT (based on years of poking around various projects, and based on the data they provide on my own projects), the data provided by GitHub are largely broken

AFAICT (based on years of poking around various projects, and based on the data they provide on my own projects), the data provided by GitHub are largely broken. Making analyses and comments on these data is futile, as I have already argued directly (and with vigour on my side) with several Redmonk employees in the past. This didn't end well, since I'm pretty passionate about these things in particular, and "the truth" in general (vs. "opinions"). I'm sure real scientists could do some interesting work on this subject, but Redmonk's methodology is anything but scientific, and in consequence, their results are just content marketing, not something that you should base business or engineering decisions upon.