Looks interesting! I saw a talk at Fluent Conf last year with similar ideas (part of it that really stuck in my mind was overlaying your org chart with commit data to see if there was a module that had people from different teams making changes in it).
Yeah, I think the techniques sound really promising. In his Ted Talk* he mentions stats like identifying 80% of bugs by looking at just 4% of the codebase, which could make code reviews and manual testing a lot more efficient. It would be interesting to see such techniques included in tooling to automate such analysis, and it wouldn't require any changes to working practices - it's just from the source code repo.
