My GitHub data challenge entry was along a similar vein. The data is stale now, but it does break it down by project and language which can form some interesting hotspots.
The globe is a really nice touch and you did a good job on canonicalization/grouping/binning.