Rolling Stone develops algorithm that identifies depression
Widely considered to be one of the greatest musical geniuses of his generation, 27-year-old Kurt Cobain committed suicide when his band, Nirvana, was at the very top of the music world. Cobain’s death anniversary is a sad reminder of how depression can put an end to young talents—and as such, it continues to raise extremely urgent questions. A quarter of a century later, it’s never been more important to ask: can we actually do something to identify, treat and prevent new suicide cases?
In 2018, a total of one million people across the world took their own lives. That’s one death every 40 seconds. For every person who commits suicide, 20 others have already attempted to do the same, according to data from Preventing Suicide: A Global Imperative, a study conducted by the World Health Organization.
In this critical scenario, different studies have suggested new ways to identify the issue and prevent tragedies from happening. For instance, a number of mental health specialists are currently focusing on the language used by people in depression, which frequently contains recurring words and expressions that could indicate a mental condition (even in its early stages). This “language of depression” can be found in everyday situations and is also present, as one would expect, in social media.
These studies have been the main source of inspiration for Algorithm of Life, a project developed by advertising agency Africa for Rolling Stone in Brazil. First, an algorithm capable of analyzing public Twitter posts and identifying an enormous variety of words, expressions and sentences that could be related to the symptoms of depression was designed. After this initial phase, a special team carefully evaluates the results so that the tool can account for context, irony, recurrence and frequency.
When a potential case of depression is identified, a Twitter profile created specifically for the project—and managed by a team of trained professionals—contacts the user through direct message and suggests that they contact a Brazil 24/7 Lifeline service.