Your cellphone knows where you’ve been. And new research shows it can take a pretty good guess at where you’re going next.
A team of British researchers has developed an algorithm that uses tracking data on people’s phones to predict where they’ll be in 24 hours. The average error: just 20 meters.
That’s far more accurate than past studies that have tried to predict people’s movements. Studies have shown that most people follow fairly consistent patterns over time, but traditional prediction algorithms have no way of accounting for breaks in the routine.
The researchers solved that problem by combining tracking data from individual participants’ phones with tracking data from their friends—i.e., other people in their mobile phonebooks. By looking at how an individual’s movements correlate with those of people they know, the team’s algorithm is able to guess when she might be headed, say, downtown for a show on a Sunday afternoon rather than staying uptown for lunch as usual.
For this innovation, the researchers—Mirco Musolesi, Manlio Domenico, and Antonio Lima of the University of Birmingham—won this year’s Nokia Mobile Data Challenge. It’s fascinating from an academic standpoint. But how exactly might it be used in the real world?
In Forbes, Parmy Olson focused on the possible law enforcement applications. If authorities could figure out what patterns of movement are associated with particular crimes, they might be able to station officers in the right place at the right time to prevent them.
The problem is that this algorithm isn’t about aggregating a bunch of anonymous data and then picking out trends. It’s about tracking specific individuals and their friends. Even if that could somehow be justified legally and morally, it isn’t practical. How would police know whom to track if they haven’t committed the crime yet?
I talked with Musolesi about his research, and he thinks a more plausible application in the near future might be commercial. Companies like Google already target you with ads based on your web-browsing habits and the contents of your email. If they could also guess where you’re likely to be at a given time, they could serve you ads relevant to that location
Or think of an opt-in service such as Groupon, which offers you discounts at local shops and restaurants based on where you live and what types of things you like to buy. If its mobile app could also start tracking and predicting your movements, it would know when to offer you a deal at a lunch spot near your office as opposed to one near your home.
The main obstacle, Musolesi acknowledges, is privacy. But cellphone companies already routinely share customers’ location data with police in case of “emergencies,” even in the absence of a warrant. (Despite widespread media coverage, the practice doesn’t seem to have drawn much outrage from the public.) And while some people might be less comfortable with the idea of commercial entities tracking their movements, others embrace it. Just look at the Foursquare users who whip out their phones to “check in” wherever they go. With a good prediction algorithm, and you’d have a reasonable chance of guessing where they’ll check in next.
For his part, Musolesi says he just finds it fascinating to uncover the hidden patterns that govern how people move about on a daily basis. No doubt plenty of advertisers feel the same way.
