How are usage stats and numbers of users collected?

I assume that the tables provided under "sources/data/usage_stats" are generated one-off statically - does anyone know how the users and tool usage stats are calculated? I find it interesting that tools can have significantly more users than usages.

Start by inspecting the tables under sources/data/usage_stats and trace how their user and tool-usage values are produced. Determine whether the data is generated statically or by a process, clarify the difference between users and usages, and document the findings.

Written by the indexing model from the issue text.