Good Reception: Using Cell Phones to Predict Behavior

Article ImageIf your cell phone knew what you were going to do at two o'clock, would that change how you planned your day? If your cell phone "predicted" correctly where you would be at a particular time of the week, how would you feel? No longer hypothetical situations, the Reality Mining experiment answers these questions.

Headed by Nathan Eagle, the Reality Mining experiment was designed to more precisely determine how people spent a period of time by tracking their daily cell phone usage. The academic research project was performed at the MIT Media Laboratory from August 2004 through January 2005. It takes advantage of the increasingly widespread use of mobile phones to provide insight into the dynamics of both individual and group behavior. By leveraging machine learning, the Reality Mining experiment is creating models that can be used to predict what a single user will do next, as well as model behavior of large organizations. Eagle and his team handed out 100 Nokia 6600 cell phones to MIT students and faculty and, using censors in the phone with information gathered from cell phone towers, were able to predict a user's location, based on phone usage over the course of six months.

The project captured communication, proximity, location, and activity information from 100 subjects at MIT over the course of the 2004-2005 academic year, which represents more than 350,000 hours (~40 years) of continuous data on human behavior. The research will examine how social networks evolve over time, the predictability of most people's lives, the flow of information, the relationship between the topology of a social network and proximity data, and how to change group interactions to promote better functioning.

According to data collected by Eagle and his staff, 35% of the subjects used their phone's clock application regularly, yet opening the clock application on the phone involved 10 keystrokes. The data showed that people who used the phone's clock application used it at their homes rather than at work. Additional results from the Reality Mining experiment show that, not surprisingly, 81% of communications from the cell phone were made via a voice message. Despite the growing popularity of text messaging, it accounted for only 5% of communications, while email communications accounted for 13%. Learning user's application routines can enable phone makers to place well-used applications in more prominent places, for example, as well as create a better model of user behavior.

In addition to his Reality Mining experiment, Eagle helped start a company called MetroSpark that "connects people who don't know each other, but probably should," says Eagle. MetroSpark is a New York-based company, which will provide a free mobile phone-based service. It will generate money from directed, opt-in, context-driven alerts to services. The data collected from the Reality Mining experiment, including how people interact with each other in social settings, provides the basis for the matchmaking algorithm for MetroSpark.

Bluetooth technology, which has been around for awhile but is just now reaching market penetration, helped pave the way for Eagle's work. "Bluetooth enables the MetroSpark service by locating people, places, and things that are proximate to the user," says Eagle. "But it's hard to comment on the staying power of Bluetooth because there are already better protocols." A Bluetooth device was initially used to allow wireless headsets or laptops to connect to a phone. Serendipitously, Bluetooth devices picked up the location of other Bluetooth devices. Eagle was able to harness Bluetooth technology by using a device inquiry scan—the device sends out a ping every five minutes to a range of ten meters. When the ping came back, it could identify not only people, but also places and things within that range. Eagle says that "anyone carrying a Bluetooth device essentially is in a ten meter bubble where they will broadcast their unique ID."

It seems as though everyone these days has a cell phone—from teens to tow truck drivers to top executives. Interestingly, Eagle's work actually would allow us to predict what possible occupations individuals have through their cell phone usage.