Everybody knows that geo-location services like FourSquare let everyone know where your favorite Starbucks or restaurant is located – that’s the point. But how about where you live?
Researchers from universities in Brazil and India analyzed data from 13.6 million Foursquare accounts and found that the user’s home town (or city) can be reliably determined from their activity on the location sharing network, including “check-ins,” “mayorships” and ”tips.”
Foursquare, a location-based social network, allows mobile phone users to post their location and add reviews, tips and other content about places they visit. Frequent visitors are designated as “mayors” of a particular location and can be entitled to particular promotions. The site has been the subject of hand-wringing over the privacy implications of sharing GPS location data. It has also been the subject of assorted hacks – notably: celebrity Ashton Kutcher’s Foursquare account was compromised by a hacker in January of this year.
The latest research is bound to add fuel to that fire.
Foursquare data could infer “with high accuracy” where the user lives. Specifically: the researchers could correctly infer the home town of the Foursquare users 78% of the time, within an accuracy of about 50 kilometers. Foursquare users, the researchers conclude, tend to accumulate Foursquare honors like mayorships close to where they live. Those awards then become indicative of the user’s home base, the researchers found.
The paper, “We Know Where You Live: Privacy Characterization of Foursquare Behavior” was presented at the 4th International Workshop on Location Based Social Networks in Pittsburgh, Pennsylvania earlier this month.(PDF of proceedings here.) The conference was held in conjunction with Ubicomp 2012 and sponsored by Microsoft Research and Carnegie Mellon University, among others.
For their work, researchers from the Universidade Federal de Minas Gerais in Brazil and Indraprastha Institute of Information Technology in India crawled Foursquare’s network and gathered public data on user accounts between August and October 2011. In particular, they collected and analyzed data on the small subset of Foursquare power users who check-in to locations on the app, leave tips for other users and accumulate mayorships (for frequent check-ins).
Operating from the assumption that Foursquare users tend to interact with venues in the city in which they live. Mayorships turned out to be the best single indicator of your location – accurately predicting your home base 52% of the time for low frequency Foursquare users and 67% of the time for higher frequency users. That was followed by tips and then “dones” or completed tasks.
No surprise, the accuracy of Foursquare improves as you zoom out. Mayorships accurately predicted the home state of frequent Foursquare users 81% of the time and the home country 93% of the time, the researchers found.