Much as I like Obama, I can't help thinking that reporting these numbers is a bad idea, because it gives them significance that they don't deserve. Online activity is not a reliable indicator of popularity. It would be amazingly easy for an unscrupulous person to rent out a little time on a bot net to get a favorable ratio... even easier than it was in 1998, when Hank the Angry Drunken Dwarf won People's Most Beautiful Person online contest.
This article does seem to illustrate the priorities of the Obama-voting demographic. The vast majority of Clinton's supporters simply don't spend their time on facebook, blogs, digg, wikis and youtube.
They should do a poll similar to this outside of a local starbucks as well. I'm sure that will be equally enlightening.
Even without that shenanigans, a candidate's popularity online seems to have little relevance to their real-world results. See Howard Dean, Ron Paul, etc.
