Add Filter for Dependabot and other Bots

[madnight]

Hi GitHub Innovationgraph Team,

It seems that your dataset includes bot pushes, which significantly impacts the ranking. I conducted some research using the public datasets from https://www.gharchive.org/, also available on Google BigQuery. By applying some filters to exclude bots (e.g., filtering out accounts with more than 10,000 pushes per month, which are likely bots), Python ranks first, and JavaScript ranks second.

[serdar-irmak]

Hi Innovationgraph Team,
Why don't you expose all the statistics like https://madnight.github.io/githut in your language chart page ?

I think Microsoft is manipulating Innovationgraph data to make c# rank higher for a special condition as below:

Each data point corresponds to the rank of a programming language based on the count of unique developers who uploaded code to a repository containing that language during a given quarter. Programming languages with a greater count of unique developers appear higher on the chart and later values appear farther to the right.

[khxu]

Thanks for reaching out (and for your patience)! We apply filters to exclude the activity of bots and other high-volume actors:

Additionally, we excluded developers whose activity volume exceeded a threshold that could be reasonably attributed to human activity and developers who were classified by our automated systems as inauthentic. Thus, Innovation Graph metrics are not useful for understanding the adoption and behavior of bots on GitHub.

Scanning https://madnight.github.io/githut/#/pushes/2024/1, that appears to plot pushes, whereas the Innovation Graph languages.csv dataset reports on unique pushers.

We'd love to report on more metrics. The limiting factor, as in many endeavors, is the cruel finiteness of time and energy. Your interest in more language-related metrics is duly noted, though, and something we'll consider as we work on additional metrics.