Leveraging the power of technology to identify cases of depression and prevent suicides

Depression is one of the most common pathologies present in human beings, and however, it is not always evident, persons with depression and suicidal tendencies tend to present behavioral patterns many times so subtle that get unnoticed by most people, and this is something normal, since not everyone has knowledge on these topics, after all, one simply can't know everything, and here is where technology can help us, cloud computing and artificial intelligence can be integrated with psychology studies, and help identify these cases.

By leveraging the power of technology you can help parents identify possible signs of depression on their children utilizing the children's social networks data.

With the data it is possible to get important metrics like how many times per week a user has posted in a social network, how possitive or negative the person has been throughout the time, association betwwen a negative mood and places visited, the many different emotions a person has through the time, like happiness, anger or sadness.

We can also get information on the most common used phrases, and map them to common words and phrases used by persons with depression.

Below you will see some sample Power BI reports consuming the analyzed data for a user's social network.

Weekly Posts

Some studies mention that if a person posts more than 3-5 times per week, that may be a possible indicator of a person having depression, of course that indicator alone mean nothing, when compared to other indicators like the emotions and negativity of posts, it begins to take meaning.

Weekly Posts

Monthly Sentiment

How positive or negative a person has been during a particular period and the trend of each can also be a possible indicator of depression. How sad and dark/obscure the person's messages have been can also be a possible indicator.

Sentiment Per Year

Sentiment Per Month

Association Between Sentiment and Places

With the data analysis it is possible to identify association between sentiment and the visited places, which is a useful key to check if when a person has been negative tends to visits a place or if the negativity tends to be a result of visting a specific place.

Sentiment And Place Association

Emotional Analysis

Furthermore, the data can be compared to case of study and even create some preddictions like when a person will possibly feel sad, or angry in the future.

Weekly Sentiment With Filters

The previous report takes advantage of Power BI filtering capabilities to help the user identify sentiment in a range, and the weeks and months where that happens. In a real case scenario a therapist or a father can search for data where sentiment is below a given threshold (45% for example),, and the report will show the months and weeks in the year when this has happened.

Posts with Sentiment and Places

The previous report takes advantage of Power BI filtering capabilities to help the user identify the posts where there is a sentiment in a given range. This report allows the user to associate the places where a sentiment is low, which can be used as additional insight to discover places that could be causing depression, or the places a person visits in order to run away from depressive emotions.

For more information about the project you can send an email to services@pticostarica.com