'Data Selfie' App Shows You Exactly Who Facebook Thinks You Are
Data Selfie follows your scrolling habits, likes, and personality.
Facebook may know you better than your friends and family do. It has catalogued your browsing habits over the years and collected information on your sexual orientation, political leanings, religious views, personality traits, and more. Now, a new plug-in can show you exactly what it thinks about you.
Created by Hang Do Thi Duc and Regina Flores Mir, second year MFA students at Parsons in the Design and Technology program, 'Data Selfie' will soon be available as both an iPhone app and a Chrome browser plug-in. The platform crawls users' Facebook activity to create a profile similar to what advertisers see, following anything the user types (even if they don't post it) as well as scrolling patterns, likes, and other interactions with the network.
The Chrome extension is still being tested and will launch in three months, but a demo video from the team shows the data that can be gleaned from your browsing habits, including levels of "agreeableness," "extraversion," and "openness."
Image: Data Selfie
Do Thi Duc, who is German but lives in New York City, said she first started developing the idea to see exactly how her browsing habits here and in her home country affected her ad profile.
"We live so much online, you start to wonder if these companies know more about ourselves than we realize," she said.
The project spread to the idea of data privacy more broadly and the increasingly advanced algorithms used to profile users online. Flores Mir said many people think they "don't have anything to hide" on social platforms, but the algorithms companies use to predict user personalities aren't as straightforward as they think. For example companies don't just mark "democrat" on your profile because you 'like' Hillary Clinton's page––the predictive algorithms are much more complex, drawing conclusions from peripheral interests and likes.
"People don't seem to understand there's not a one-to-one correlation of the things that you do online, it's the seemingly mundane things you do online that are predictive of who you are," she said. "That is what Data Selfie is trying to show."
The team is currently using a machine-learning system from IBM Watson to develop the technology, but they are working to create their own in the future through partnerships with other researchers. The market they are targeting is parents and teachers, allowing children and teens to learn exactly what information companies have about them and to better understand their internet consumption. The team has spoken with advocacy organizations like EFF about how to target users and help them better understand privacy.
"It's not about watching them ore controlling them, just educating them about how the internet works so they can make decisions themselves," she said.