Computer scientists are analyzing Twitter tweets to gather key information on the prevalence of common mental illnesses.
Researchers at Johns Hopkins University in Baltimore say their new computer program can sift through volumes of publicly available postings on the social media website, and detect certain ‘language cues’ associated with particular disorders, such as depression, post-traumatic stress disorder, bipolar disorder and season affective disorder (SAD).
One benefit of mining relevant mental health data in Twitter posts is that analysis of the information can be delivered to medical professionals much quicker and cheaper than with current, traditional methods.
The data on mental illness trends discovered during a Twitter search can even provide information for specific geographical areas, which would be handy for public health officials and medical providers during times that follow natural and man-made disasters.
The Johns Hopkins scientists evaluated over eight billion tweets in developing their computer algorithms that look for specific words and language patterns in the Tweets. For example, if information regarding disorders such as anxiety or insomnia is desired, the algorithm would pour through the tweets and look for words and phrases such as “I really don’t want to get out of bed today” or “I’m feeling really sad today”.
This new Tweet-based data gathering and analysis system for mental illness was built off a similar system that Johns Hopkins researchers developed back in 2013. That system filtered out irrelevant online chatter to produce real-time data on cases of influenza.
As the researchers put their new system to the test it revealed a prevalence of Post-Traumatic Stress Disorder (PTSD) among military personnel at US armed services installations that regularly deployed combat troops to Afghanistan and Iraq. It also detected Tweets indicating higher than normal symptoms of depression in areas where unemployment was high.
“Using Twitter to get a fix on mental health cases could be very helpful to health practitioners and governmental officials who need to decide where counseling and other care is needed most,” says Mark Dredze, an assistant research professor in the Whiting School of Engineering’s Department of Computer Science.
“It could point to places where many veterans may be experiencing PTSD, for example, or to towns where people have been traumatized by a shooting spree or widespread tornado damage,” Dredze says.
Privacy issues were of the utmost concern to the researchers as they developed their new mental health analysis system. They noted that information on mental health issues gathered and analyzed by their new tool does not reveal the names of people who publicly tweeted about their disorders.