New data from a new study by researchers at Columbia University and Google.
Stripping a big data set of names and personal details is no guarantee of privacy. Previous research has shown that individual shoppers, Netflix subscribers and even taxicab riders are identifiable in heaps of supposedly anonymous data.
Now, a team of computer science researchers at Columbia University and Google has identified new privacy concerns by demonstrating that geotagged posts on just two social media apps are enough to link accounts held by the same person. The team
willpresent[ed] its results at the World Wide Web conference in Montreal on April 14.
The team developed an algorithm that compares geotagged posts on Twitter with posts on Instagram and Foursquare to link accounts held by the same person. It works by calculating the probability that one person posting at a given time and place could also be posting in a second app, at another time and place. The Columbia team found that the algorithm can also identify shoppers by matching anonymous credit card purchases against logs of mobile phones pinging the nearest cell tower. This method, they found, outperforms other matching algorithms applied to the same data sets.
A New App
Location tracking is now embedded in phones and many apps precisely because it’s so useful. It’s what allows you to get directions from here to there, learn that a friend is unexpectedly nearby, or that a store in the neighborhood is offering a promotion. These perks, however, come with large privacy risks that remain poorly understood.
One problem with the current system is how opaque it is to the people unknowingly leaking their data simply by carrying around a phone. With two Columbia undergraduates, Danny Echikson and Stephanie Huang, Riederer built a related tool, You Are Where You Go, to let individuals audit their social media trail.
With a few clicks, the tool retraces your steps on Twitter, Instagram and Foursquare. A few simple algorithms process this information to make relatively accurate inferences about your age, ethnicity, income, and whether you have kids.
Read the Complete Article
Direct to Full Text of Research Paper Discussed in Article, “Linking Users Across Domains with Location Data: Theory and Validation” (13 pages; PDF)