Advertisement

Twitter feed can help predict depression, researchers find

A portrait of the Twitter logo in Ventura, California December 21, 2013. REUTERS/Eric Thayer

Reading through someone’s Twitter feed can quickly tell you what they ate for lunch and what they thought of last night’s Master Chef, but one study thinks it can also tell you if they are depressed.

Research is currently underway at Microsoft Research Redmond into how an automated system could be used to scan Twitter feeds for signs of depression.

Eric Horvitz, co-director of the group, told TIME that he thinks the technology is possible, and they’re working to make it happen.

“We wondered if we could actually build measures that might be able to detect if someone is severely depressed, just in publicly posted media. What are people telling the world in public spaces?” asks Horvitz. “You might imagine tools that could make people aware of a swing in mood, even before they can feel it themselves.”

So far, Horvitz and his team have developed a model that can scan tweets and predict depression in Twitter users with a claimed 70 per cent accuracy. They’re still working on increasing the accuracy, so the 30 per cent who will become depressed don’t get missed and so they don’t return as many “false positive” depressive Twitter users (currently about 10 per cent of cases fall into this category), but the evolution of the technology is certainly promising.

[ Related: Let’s talk: New site collects the million reasons people have to live ]

After looking at 2.2 million tweets from 476 Twitter users, the researchers found certain signs that indicated a user was likely depressed. Obvious signs, like language about sadness or avoiding depressing situations, were generally clear indicators from the depressive group. There were also less obvious signs, though, like the time a person tweeted (depressed people were more likely to tweet late at night) and how often they interacted with others (depressed people were less likely to have exchanges on Twitter with others).

The technology is still a work in progress, but Horvitz says he is hopeful that it could help stop some of the 30,000 suicides that take place in the U.S. due to depression each year.

“If we can even save through interventions a few of those 30,000 people each year, it will make this research well worth it.”

Need to know what’s hot in tech? Follow @YRightClick on Twitter!