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.
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.
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.
Didn't Howard Dean beat John Kerry in 2004 online?