[UPDATE] I cleaned up the source a bit, so it’s suitable for human consumption.
I was adding more bells and whistles to my mail system — this time, a command-line program to parse through SpamAssassin’s configs and tell you about rules, and one to add headers as such — when I realized that all that data I was basing the pointless little graphs in my earlier post on was a little rough, and SpamAssassin already has the data on what I consider ham and spam. D’oh!
So, here’s another one, based on sa-learn --dump magic. (Here’s source.)
