Tuesday, June 9, 2009

Reality Mining

A few years ago, MIT media arts professor Alex Pentland distributed one hundred Nokia cell phones to students and faculty. Each was equipped with software that let Pentland’s team map the social interactions of the people carrying the phones. They tracked some 130,000 interactions, including calls, text messages and the proximity of the devices to each other, as measured by Bluetooth.

The goal: to learn about the behaviors of social networks and the types of people who influence their dynamics the most. “We weren’t able to measure this stuff at such a large scale before smart phones,” says Pentland, the author of the recent book, Honest Signals, and the chief of the Human Dynamics Lab at MIT. “By using data from mobile phones, electronic ID badges, or digital media—what I call ‘honest signals’—we can create a God’s-eye view of how people in organizations interact,” Pentland says. "We can even see the rhythms of interaction for everyone in a city.” Mining mobile data this way would be useful, he says, for tracking such things as flu outbreaks and traffic patterns. (See video here)

Pentland calls this new field of dynamic data “reality mining.” Along with Columbia University Chief Scientist Tony Jebara and a half-dozen other location-awareness technologists and researchers, Pentland co-founded Manhattan-based startup Sense Networks a couple of years ago to make commercial use of the data: the company’s Citysense application for iPhones and BlackBerries is one of the company’s first commercial applications of the research. For now, it’s a free download that shows which areas of San Francisco are the No. 1 hotspots for nightlife—where most people are going or gathering at any given time.

“Half of the people in the world have cell phones and are creating data, tons of it now,” says Sense CEO and Co-founder Greg Skibiski. “This data, in a few years, will be the lingua franca of the entire planet about how people move around globally. We believe location data is the best descriptor of who is a human being.” Adds Pentland: “The motion detectors in your iPhone, for example, know if you’re sitting or walking. Web 2.0 stuff is nice but reality-mining is different.”

To be sure, reality mining raises red flags among privacy advocates. Writer Howard Rheingold says: "The same technology that could let you know if a good Chinese restaurant or old friend is in the vicinity could also betray your location to a totalitarian government, neighborhood spammers, and your vindictive ex-spouse.”

But that's why Pentland and Sense Networks are now pushing what they call a New Deal on Data —an informal privacy pact for the mobile Net set that Pentland unveiled at the World Economic Forum in Davos earlier this year. “People always say, what about my privacy? But what if privacy can go beyond that? Pentland queries. "Like what about data ownership?” The New Deal on Data, Skibiski says, is based on old English common law tenets that give indivdiuals the right of poessession, use, and disposal. "If you’re a company and you're keeping people's location data, your customers should own their own data and have a say over whether they should be destroyed. It's something we all [in the geo-location industry] should be working towards.”

But are they? So far, says Skibiski, MySpace adheres to it and Facebook is considering it. Nokia, which is working with movie theaters in geo-location experiments to measure theater attendance patterns, also supports the proposal. “There’s a need for some guidance here, absolutely, so as to protect against abuse—but most companies get it, I think," says McDowell, who is leading a Nokia experiment for Remax, the real estate company, to enable people to point their iPhones at a house that’s for sale and get information about its dimensions. “I don’t want to know if you’re visiting your girlfriend," she says. "I just want to know if you are driving down the highway at a certain speed so I can predict traffic flows in a region.”

For more on the new world of geo-location services, check out Mathew Honan’s recent take for Wired on the promise and the perils of context-aware devices.

—Marcia Stepanek

(This post first appeared on PopTech and is being reposted here with permission)

(Illustration by AskinTulayOver for istock.com)

Labels: , , , , , , , , , , , ,


Post a Comment

Subscribe to Post Comments [Atom]

<< Home