Surely this is too small of a sample to be accurate of Twitter users? http://bit.ly/K085n (via @guardiantech)