It seems like absolute count of occurrences, not normalized to codebase size.

Even more informative would be to plot the occurance rate within new code.

OKR for H2: Increase edginess of Linux for a less corporate feel

Key result: Boost occurrence of swearwords by 20%

Key result: Create a new metric that tracks relative swearword use per line YoY

Key result: Attract at least 100 comments on HN or Reddit about the new code

Comment mastery