Mobile devices can track where you are, where you have been and soon may be able to predict where you are going. Researchers in the U.K. have created an algorithm that can predict where smartphone users will go, even before they get there, Technology Review reported. The tech tracks a user’s mobility patterns and adjusts for anomalies by factoring in the patterns of friends as well as friends of friends. The method was found to be remarkably accurate and on average was less than 20 meters off when predicting a user’s location 24 hours in advance. When the algorithm didn’t take into consideration the previous location of friends and mutual friends, it was found to be an average of 1,000 meters off when predicting a user’s future location.
The researchers cautioned, however, that the 200 participants were all students and lived within 30 miles of one another, and their movements were fairly predictable to begin with. Mirco Musolesi, a computer scientist at the University of Birmingham who led the study, said the findings were remarkable nonetheless because “we are essentially exploiting the synchronized rhythm of the city” for greater predictive insights.
The ability to predict a user’s future location would be highly valuable and could allow for more accurate recommendations and better targeted advertising systems.