Research shows suicidal tweets reflect state suicide rates

322

Social media has boomed in recent years as a way for friends to connect across distances, but a new study by a group of BYU professors and student research assistants shows that it may also be used to detect individuals at-risk for suicide.

Computer science professor Christophe Giraud-Carrier, along with BYU health scientists Josh West and Carl Hanson, recently published their work in the academic journal Crisis detailing how data mining was used to sort suicidal tweets.

“We looked at suicide-related tweets and plotted them against actual rates,” co-author Josh West said. “We found that tweets coincide in terms of frequency and relevance with actual suicide rates.”

Computer science professor Christophe Giraud-Carrier, along with a team of six other researchers, found that suicidal talk on twitter mirrored actual suicide rates. This technology could potentially be used as an early-warning method.  Photo courtesy BYU Photo
Computer science professor Christophe Giraud-Carrier, along with a team of six other researchers, found that suicidal talk on Twitter mirrored actual suicide rates. This technology could potentially be used as an early-warning method.
Photo courtesy BYU Photo

The team studied tweets from all 50 states over a three-month period, sorting out conversations involving suicide as well as potential risk factors, such as bullying. The research yielded 37,717 troubling tweets from 28,088 unique users, according to a press release.

Co-author and computer scientist Christophe Giraud-Carrier explained that he hopes people will look at these findings and recognize that there is value in what happens in social media because it often reflects real life.

“With social media, kids sometimes say things that they aren’t saying out loud to an adult or friend in person,” Giraud-Carrier said.

The researchers suggested that these findings may serve as a helpful addition to current suicide-prevention efforts.

“We spend lots of money in the U.S. to gather data for health research purposes,” said West. “(With this technology) we can potentially catch things earlier and we can replicate expensive findings virtually for free at a comparable quality.”

Giraud-Carrier and the Data Mining Lab have been collaborating with the Department of Health Sciences for some time, researching relevant health issues through social media such as alcoholism and Adderall abuse, in addition to this most recent research on suicide.

Giraud-Carrier explained that this new research on suicide rates has led to the development of a new app for Facebook that will monitor the moods of a user’s friends and then alert these users if their friends are showing warning signs.

Though the app is still in the prototype stages, Giraud-Carrier said that he hopes that people will be able to utilize the app to take action when their friends are flagged for being at-risk and either help them or refer them to a professional who can.

“Suicide is preventable,” said co-author and health scientist Carl Hanson. “If we can recognize through social media those who are at-risk … we could potentially provide an avenue for intervention.”

 

Print Friendly, PDF & Email