For reasons that I've not quite figured out, for the last several weeks, I've felt all but compelled to map article links in wikipedia. I think it would be interesting to (attempt) to see how all of the pieces fit together. I mean, they post XML copies of all of the articles at least once a month. All I need to do is download a 5.7Gb compressed en.wikipedia.org mirror, decompress the file, parse it and locate some kind of graphing software that won't implode the universe when it sees the size of the data being passed to it...
Something tells me this is going to take a LONG time... until the last step which I fully expect to be infeasible. Either way, I think this project can safely wait until summer.