First get the commit message history into a text file (the last commit is r287):
svn log >> test.txt
Then with a bit of VIM trick and manual step, remove unwanted lines:
Select of all the text, and import into
Tadaaa, here comes the result :)
Indeed, we are data scientists here in atmospheric science!
Alternative wordcloud implementation comes from Andreas Mueller. I saw his A Wordcloud in Python titled posting via Planet SciPy. In order to run his wordcloud.py I needed to change unsigned int[:,:] integral_image -> unsigned long[:,:] integral_image in query_integral_image.pyx to get past 'ValueError: Buffer dtype mismatch, expected 'unsigned int' but got 'unsigned long''.
Then feeding the same log data I get the following image, which is nicer looking than the wordle's image :)