With Lexicon we, the users of these mass social networking (read preference collection) sites, have the ability to mine some of the data we provide. Facebook, even give the data a first-cut cleaning to correct for common mispellings (sp.) and problems with apostrophes. Notwithstanding the absence of a scale on the y-axis, the conclusions we draw from the data are not always incredibly interesting. Take, for example, the graph below showing the appearance of each candidate's name on Facebook walls. You can add the word 'delegates' to your search and watch the peaks line up. You can pick out Super Tuesday (Feb 5) easily. More than that, you can see that Obama is getting a lot more free press out there in the Facebook-a-sphere.
I think this could be useful to get a rough idea about some associations we suspect to exist. Unfortunately the wall is not the best place to collect data as only certain topics are discussed and the makeup of the users is likely to screw up any generalizablity, and sometimes the scale of the two things you are comparing makes it difficult to find what you might be looking for (see 'cold' vs. 'bike'). Nonetheless it is fun to see you suspicions confirmed (no one mentions 'smog' in Dec or Jan)and to think a bit about what people are talking about. Remember that you will not be able to infer causation from anything you find here so be humble with what you discover. Here are a few more of my favorites.