Twitter may have an unexpected new purpose: predicting disease. A study, published in the journal Psychological Science, found that language—whether positive or negative—could be an accurate forecaster of heart disease rates.
Researchers at University of Pennsylvania analyzed language in 1,300 U.S. counties, and found that a Twitter feed full of anger, stress, or fatigue correlated to a higher risk of heart disease in that county, and vice versa. When compared with traditional predictors—like income, education, or even weight—Twitter feeds proved to be even more accurate.
And when compared to a map of actual deaths from heart disease, the map of Twitter-predicted death rates was incredibly similar.
The study wasn’t meant to focus on one person’s risk of heart disease, but rather how a community as a whole can predict the health of its individuals. Specifically, they found that “angry tweeters” weren’t necessarily suffering heart attacks themselves, but “angry communities” saw a higher risk of disease. This lines up with recent research from the University of Michigan, in which scientists found that cohesive neighborhoods saw a 22 percent decreased risk of heart attack.
“Twitter seems to capture a lot of the same information that you get from health and demographic indicators, but it also adds something extra,” researcher Gregory Park said in a press release. “So predictions from Twitter can actually be more accurate than using a set of traditional variables.”