Glen Coppersmith, one of a number of computer scientists at John Hopkins University (JHU) said that looking at tweets from people who publicly mentioned their diagnosis lets them speedily and cheaply collect data on seasonal affective disorder, depression, bipolar disorders and post traumatic stress disorder (PTSD).
The scientists trawl through tweets and use computer technology to counter the high costs of collecting mental health data using surveys.
“With many physical illnesses… there are lots of quantifiable facts and figures that can be used to study things like jow often and where the disease is occurring. But it’s much tougher and more time consuming to collect this kind of data about mental illnesses because the underlying causes are so complex and because there is a long standing stigma that makes even talking about the subject all but taboo,” said Coppersmith.
Coppersmith also said the team didn’t want to replace surveys to track trends in mental illness. Its techniques are meant to complement them.
PTSD is more prevelant at military installations in Iraq and Afghanistan. Depression happens in places that have high unemployment rates.
So what of the algorithms? The scientists look for words and language patterns including phrases like “I just don’t want to get out of bed”. The scientists looked at eight billion tweets.