> > The ranking stuff is always interesting... but it always seems to be a > solution in search of a problem. The problem usually seems to be, "Who > should I follow?" and then the rankings seem so obvious as to be low > value... and I start wondering how this could be personalized. The big > rankings tell me who *everybody* should follow, but I want to know who *I* > should follow... and now I'm wondering how, with all the APIs around, I can > base that on much more than just Twitter-sourced data.
Yes... agreed. This is why I was thinking about creating vertical rankings. Technology, politics, entertainment, etc. I think we'll roll this out in the next big push.... Trying to see how much value there is here for the community though. Using something like wefollow as the source might be interesting :) > I'm rambling... but it was good to suddenly realize that this burtonater is > the same guy I've been paying attention to over the years. I'll be up in SF > next week if you'd like to connect. Clustering might also be interesting... take all the tweets, cluster them, then rank the users who create those tweets. That's easier said than done though... lots of data and computational time.... We're in SOMA if you want to grab coffee.
