The Facebook Likes are one of the easiest way to interact with a huge number of sites, and users have loved the feature ever since it came out. What most of them don’t suspect is that meaningful data about their personality traits can be extracted from something as simple as a series of Likes. In the Big Data world, people often say “more data is better than smarter algorithms” and this seems like a case where this is true again. Researchers at Cambridge’s Psychometrics Centre and Microsoft have mined the Like usage of 58,000 users and were able to extract accurate information about their religious affiliation (82% accurate), race (95% accurate), divorced parents (60% accurate) and even substance abuse (65%-73% accurate). These are not things that one may think could be obviously derived from a series of Likes, but the study proves that it’s possible.
And this is only one simple source of data. Imagine what could be done with more access to your digital trail online. At the moment, it’s not hard to guess that this information would be used for marketing purposes, but when it comes to religious and political affiliation, you can imagine what repressive governments around the world could do with that information. Researchers believe that unless there is transparency about how the data us being mined, users may be deterred from using services that may record this type of trail. “This study shows, in a scary way, how transparent users have become on the Web. At the same time, it’s alarming how willingly users share their personal data without second-guessing how their personal data could be used.” says Secure.me and they are right: people will probably not care for a while. What’s your take on this? What is your tolerance to data-mining and when would it start to be a real issue for you?