correct me if i’m wrong

Check out my flock on TwitterSheep. This simple service aggregates the words used in the bios of followers for a specificed user. I am surprised that all of these keywords are generated solely from the bios–some seem rather unlikely, like “salsa” or “techcrunch50.” Lo and behold, I could quickly find these friends. It’s interesting to see the collective characteristics of your followers, and it certainly highlights the idea that you are defined by your friends, which is particularly emphasized in the social networking age.

A more interesting idea would be to do a similar aggregation of the actual tweets rather than bios. This would be slightly more difficult, because of the high amount of noise that would be produced. Simple word sense disambiguation or entity extraction could help produce a quality tag cloud. The computational linguists have some work to do on Twitter, such as Happy Tweets and what I hope are more sophisticated analyses tools in the future.