This is very cool. It should be a lot easier to design new tools like git-flow if you have hard data on how people use Git. It'd also be interesting to compare

This is very cool. It should be a lot easier to design new tools like git-flow if you have hard data on how people use Git. It'd also be interesting to compare how Git is typically used compared to, say, Mercurial or even Subversion, in order to quantify criticisms people might have of other tools.