I don't think this is terribly meaningful. Do you take into account unmoderated communities? Some communities and mods are also more ban happy than others, so one instance can have communities that very rarely ban and ones that ban a lot, and how big those communities are will also vary. A more meaningful analysis would try to measure the impact of ban-happy communities by adjusting for their size/activity or would compare individual communities. Edit: Some communities or mods also get harassed a lot and therefore need to be more ban happy (like womens stuff), but I don't think accounting for that would be within the scope of what you're looking at, but it's worth being aware of.