Study: It Takes Just 5 Tweets to Figure Out Where You Live

We see you.

Illustration: Jose-Luis Olivares/MIT

Much of the discussion about the loss of privacy in modern life comes with the assumption that maybe the government or a global corporation is tracking our activity. We can gripe, but a lot of the damage is self-inflicted with check-ins on Facebook and Instagram photos with locations attached. However, a new study has found that one big piece of information — where you sleep — can be inadvertently revealed in just five tweets.

Researchers at MIT and Oxford University found that normal people can identify where a Twitter user lives and works with 85 percent accuracy just by looking at five days of their tweets.

“It is extremely simple for people with very little technical knowledge to find out where you work or live.”

“Many people have this idea that only machine-learning techniques can discover interesting patterns in location data,” Ilaria Liccardi, a research scientist at MIT’s Internet Policy Research Initiative, told MIT News. “And they feel secure that not everyone has the technical knowledge to do that. With this study, what we wanted to show is that when you send location data as a secondary piece of information, it is extremely simple for people with very little technical knowledge to find out where you work or live.”

The study certainly stresses that turning on Twitter’s location-reporting service opens a range of possibilities that may not be immediately clear. Even a single day’s worth of tweets with location information enabled can reliably point observers to your home and workplace. Simply placing the Twitter locations on an online map makes the process a lot easier, too. When given a single day’s worth of tweet-location information on a map, study participants were able to identify the users’ homes with 65 percent accuracy and their workplaces 75 percent of the time.

Bizarrely, the participants did worse when given three days of information than a single day, though the performances were best with five days of data. The researchers attribute the mix-up with three days of data to there being just enough information to be confusing without enough to provide clarity.

“We want to investigate that,” Liccardi says. “When we asked participants ‘Which amount of data do you prefer?’ most of them said ‘medium,’ even though it was the one that they got the least right. So you never know about perceptions.”

A few tweets here and there pinned to your location may seem innocuous, but those who are weary of granting the entire world detailed information about their homes or workplaces should take note of this study. After all, if a regular person only needs five days of information at most to track down your home and workplace, imagine what a machine analyzing your entire Twitter and online history can learn about your life.