During the last 50 years, Rolling Stone became a witness of our times. It saw one of the greatest legends of rock, Kurt Cobain, write some of the most iconic songs of all time. Chords that will live forever in everyone’s mind. But no one really understood what was behind those lyrics. Neither did us. Actually, academic studies show that Nirvana used, frequently, terms of the depression grammar in the lyrics. It means that he was asking for help, but nobody listened to it. 25 years after Kurt killed himself, requests for help keep being written and unnoticed. This time, on social media.

According to World Health Organization, suicide gets a person killed every 40 seconds. It’s almost 800,000 people per year, more victims than war, cancer and urban violence. And still according to WHO, when identified in advance, 90% of cases can be avoided.

Recent academic studies have proved that some words and sentences used frequently are directly related to depression feelings. It’s the depression grammar. Those studies made us realize that Kurt was asking for help through his lyrics, using, frequently, those terms.

25 years after Kurt’s death, people suffering with depression keep writing requests for help. But now, everything is more clear, and we can help.

In partnership with Brazilian Association of Psychiatry, we created an algorithm that tracks and analyzes public posts on Twitter looking for words and phrases of the depression grammar. According to frequency, recurrence and the strength of the term, it can define in which stage of depression each person may be. In other words, the algorithm is like digital hears for people shouting for help. Once a serious threat is identified, the user receive a direct message suggesting to call CVV, Brazilian Lifeline service for help.

There are already several academic studies and data about the depression grammar. The problem is that all this information alone doesn't help prevent suicide. We needed to scale up these data so we would be able to identify and prevent new suicide cases. And, although technology and social networks are often portrayed as a disease of our time and the cause of so much depression, we saw them not as part of the problem, but rather as an important part of the solution. To develop this idea, we used machine learning to train our algorithm so that it was able to recognize and understand words, expressions, and also the emotions behind posts on Twitter - one of the most used social networks of Brazil’s youth. With all these behavioral insights, the algorithm could understand the stage of depression of each one and also the best way to help them.

There are already several academic studies and data about the depression grammar. The problem is that all this information alone doesn't help prevent suicide. We needed to scale up these data so we would be able to identify and prevent new suicide cases. And, although technology and social networks are often portrayed as a disease of our time and the cause of so much depression, we saw them not as part of the problem, but rather as an important part of the solution. To develop this idea, we used machine learning to train our algorithm so that it was able to recognize and understand words, expressions, and also the emotions behind posts on Twitter - one of the most used social networks of Brazil’s youth. With all these behavioral insights, the algorithm could understand the stage of depression of each one and also the best way to help them.

We needed to launch the Algorithm of Life on a landmark date for both Rolling Stone and the problem. And Kurt Cobain's 25th death anniversary was a sad, yet powerful reminder of how depression can put an end to young lives. It was an opportunity to put the subject on the stage without huge investments on media. With only a few ads on social media and websites, we were able to draw the attention of the media, experts and - especially - young people.

After the release, the project continued to gain awareness and to reach more and more people organically: young people sharing the Algorithm of Life with their friends and also parents becoming aware about depression signs. This all kept the project alive.

The Algorithm of Life broke the silence and the taboo around depression and suicide, making people discuss the problem and the depression signs, on social media, on the most watched program on Brazilian TV, and even in universities – where the Algorithm of Life inspired symposiums. The project reached 12.8 million people.

But, most importantly many people identified by the algorithm answered our approach thanking for our help. It proves the efficiency of the tool. We don't know exactly how many lives we saved, but if it helped one single person it was already worth it.