OK, I lied, it's not too exciting. I implemented the interquartile range method first (because it was easier than calculating the standard deviation) and ran it on a set of markers. It found a lot of markers that were in the correct city, but were definitely outliers from a statistics perspective. These markers are all dental clinics (more or less), and it turns out they're not evenly distributed throughout a city. So I kept the interquartile range method but added a test to only reject it if the distance from the median point was more than .25 latitude or longitude. This got rid of the false positives (and kept something that indeed wasn't categorized correctly!), so that's what I'm using. The sad part is that I'm pretty sure that that would give the same result if I just did that test and not the interquartile range, but I don't have the heart to test that. So, not so interesting problem after all, but at least it's presumably solved.
I'm definitely making more progress on this side project, and hopefully I'll work on it tonight a bit and maybe be "done" for now ("done" because it's never done, stuff keeps getting added, which I'm going to have to deal with at some point. ugh...) On the plus side, I'm writing a lot of stuff that can at least be theoretically reused for some other project. *shrug*
Here's a good article about lobbying to get your industry millions. Ugh. Hopefully the fact that the article was written will shed some light on the system. I am proud, however, that my former congressman Lloyd Doggett was mentioned as trying to end the offending credit. Good for him!
Some indie rockers are refusing to license their songs for Hummer commercials, turning down big money in the process. Good for them!
Wow, this is turning into Link Tuesday. One more: here are the local commercials that Rudy's aired during the Super Bowl. Good stuff!