I am writing subverion statistics graph generation utility named SVNPlot. It is inspired by the graphs generated by StatSVN. However, the SVNPlot graph generation is in two steps (a) first it creates a sqlite3 database from the subversion log information (b) actual graphs are then generated by extracting the database sqlite database (using simple sql queries).
I think using sql to extract the graph data from the log information is resulting in greater flexibility and good performance. Right now the SVNPlot only generates graphs but it very easy to extract any other stats from the generated sqlite database.
SVNPlot is written in python and it uses excellent Matplotlib package to generate the graphs. The code is available on SVNPlot page on google code (license is New BSD license). The sample graphs generated for Rietveld repository are available at http://thinkingcraftsman.in/projects/svnplot/index.htm