Mobile Phones Reveal the Behavior of Places and People

Tony Jebara, Chief Scientist, Columbia University & Sense Networks

Date: Wednesday, March 4

Time: 12:15 - 12:35 PM

Location: Salon E

As more of us generate GPS data with our mobile phones, how can this aggregated information give us an unprecedented new understanding of the people, places, and rhythms that make up our cities? How can we index the real world using location data? Location data combined with learning algorithms lets us cluster different places and people into social categories and tribes. By harnessing this rich, natural and anonymized data, unprecedented possibilities emerge for user modeling, marketing, advertising, recommendation, search and collaborative filtering. Using machine learning algorithms, we can infer the context of a place and the tribe of a user from just their location data. It turns out that the flow and movement of people through the city (who is where and at what time) defines places and their character. Similarly, a person's movement trail through the city reveals their personality and tribe. With location data, we build a network of places (how similar is place A to place B) and a network of people (how similar is person X to person Y). These networks let us cluster places and people as well as compute next-generation demographics and analytics. As your cell phone learns about you, it helps you find people, places and things you are interested in and your phone's mapping software becomes your personal social navigator.

Leave a comment