, a project built by a bunch of computer science nerds from Carnegie Mellon University, aims to provide some human-level insight about how neighborhoods in San Francisco (and New York and Pittsburgh) organize themselves. By letting their fancy-sounding machine learning algorithm analyze tweets and Foursquare check-ins, their system attempts to reveal dynamic neighborhoods that exist outside of made-up real estate boundaries, based on the places locals like to frequent. Except, we already knew all this stuff.

Near SFist's Western Branch Office, for example, we learn that patrons of the white-hot Divisadero corridor occasionally like to venture north for a screening at the Sundance Kabuki or a couple drinks in the Lower Haight. Lower Haighters, by the way, apparently don't ever leave a 1-block radius.

Meanwhile, across town in South Beach, Carnegie Mellon's hivemind discovered the top five "unique" things to do are: "Baseball Stadium, Brewery, Tech Startup, New American Restaurant and Sports Bar." Hardly a groundbreaking insight there, and supposedly the data will get better as more people check in, but it's still a fun toy to poke around if you're one of those nerds that gets off on things like interactive maps and data visualizations. Which you can do right here.

Via: Fast Co.
(Thanks for the tip, Jackson)