Visualising Change In Version Controlled Files

Question

We have been experimenting with using data visualisation techniques inspired by Edward Tufte to display our test suite and it has been very effective.

I would like to extend this to our Subversion Repository as I feel that there is a lot of information buried in the commit history that COULD be better represented in a graphical format.

I would like to be able to identify at a glance things like:

• which modules are comparatively stable - a lot of writing - a little maintenance and which ones have been written and rewritten
• which modules are all one persons work and which are the work of many

Ideally I would like to annotate this information with other stuff from testing and performance tools, like:

• code coverage
• xref stuff like function call graph
• mebbies even things like processor utilisation under consistent load

Anybody good any good tips, examples, utilities, etc, etc...

Our shop uses mostly the mighty Erlang but we will take heart and inspiration from any source.

Answer 1

Check out StatSVN as an example of a Subversion statistics generator:

http://www.statsvn.org/

http://www.statsvn.org/demo/ruby/

Answer 2

You probably have seen codeswarm which made some headlines earlier this year when it was used to generate some cool videos of collaboration in Ruby on Rails--see the Visualizing Rails & Git blog post for a great summary and sample videos.

You might also get some ideas from history flow, which Jeff Atwood linked to in a recent Coding Horror post.

Answer 3

You can try SVNPlot. It first creates a local sqlite data from the svn commit log messages. Then it uses sql queries and matplotlib to generate various graphs from it.

You can use it the sqlite database to add your custom queries and additional graphs.

(Disclaimer - I am main author of SVNPlot. Do let me know if you find it useful or if you have any suggestions on improvements)