Imagine if public health workers could monitor social media to take the emotional temperature of a specific population and use that data to predict epidemics.
Researchers from the Department of Energy’s Pacific Northwest National Laboratory (PNNL), in Richland, Wash., had that scenario in mind when they conducted a new study to learn more about the correlation between emotions expressed on Twitter and flu outbreaks. Their findings were published in the journal EPJ Data Science.
They looked at 171 million tweets from users located near U.S. military bases in 25 domestic and six international locations. They studied tweets from military personnel and a control group of civilians. Then they compared the data to the number of visits to military health clinics for flu-like symptoms. The researchers wanted to find out if the opinions and emotions expressed could have any correlation to doctors’ visits for illnesses like the flu.
The study revealed that during periods of high clinic visits for flu-like symptoms, Twitter users expressed sadness or neutral sentiments more frequently. Periods of low clinic visits for flu-like symptoms correlated with Twitter users expressing more positive sentiments, anger, and surprise. One possible explanation is sick people have lower energy and that translates to less visceral, more neutral or melancholy tweets. But there is still more research to be done before the study’s authors draw those kinds of conclusions.
Researchers chose to focus on populations around military bases as those “communities present a unique case study because they are semi-closed populations of people who share common location, responsibilities, and way of life. They actively use social media to stay connected with their unit at home or with friends and family when deployed.”
Svitlana Volkova, a PNNL data scientist and lead author on the study, hopes that by studying the psychological state of Twitter users, public health officials will eventually be able to predict and track epidemics in real time.
Some health workers already use technology, like Google Flu Trends, to track spikes in searches for specific, illness-related keywords. But Volkova’s research could unlock broader data sets by tracking the emotional state of a community, not just specific language.
“Opinions and emotions are present in every tweet, regardless of whether the user is talking about their health,” she said in a statement. “Like a digital heartbeat, we’re finding how changes in this behavior relate to health trends in a community.”
Now that the preliminary research is done, Volkova and her team will study whether the emotions identified in the study can be used to predict a change in health trends before it happens.
“If this method works in real time,” PNNL says, “public health workers could look into the future by asking ‘How are your tweets feeling?'”