I have just finished reading Freakonomics. It’s a great book, but it’s gone too fast! It took me about 3 days to complete it. Also, the authors are right to indicate early on that there is no real unifying topic. Maybe a longer discussion on how they obtained those data would have been useful.
Anyway, it just happens that I read recently about Mass Observation in the New Yorker. It strikes me that the data collected by MO is exactly the kind of data that Steven Levitt would use. I wonder if he has already consider it? Or maybe the data are simply not formalized enough to be useful.