I recently heard ~80% of empirical data used for breakthrough social science research will use mobility contexts and sensing to create new iterations of theory - I can't remember where I heard that. To lend some credibility to that statement, head over to Orange's UrbanMobs project showcase. The time-lapses are amazing.
I'm thrilled to see an Operator harvesting location hand-off, transactions, and traffic types to build spatiotemporal models of urban network activity. I've often thought they were better suited to create a platform for computational social science. Next, how cool would it be if Operators mashed up the data to build geographically filtered keyword Search for a real-time Web of text messaging.