Social Media Posts Can Predict Mental Illness

We are all aware that marketing firms collect huge amounts of data to develop ad campaigns and to identify target audiences for marketing purposes. They use massive amounts of data in search of patterns and trends to assist them in creating and perfecting their products and services. Collecting and analyzing data has become a common practice worldwide.

There are growing ethical concerns regarding how this massive amount of information is collected and then how it is being used. The general public is painfully aware that our information is bought, sold and traded for a variety of reasons—some nefarious and wicked.  As our concerns continue to mount and are compounded by reports of cyber criminals robbing these information warehouses of enormous amounts of valuable data; it is refreshing to know that there are those who seek to use this information for good and not evil.

The same way marketers and big business have been gathering and analyzing data—so to have governments, scientists and law enforcement agencies. These entities harvest and collect data for a variety of reasons. 

Large amounts of personal information are collected and used in the prediction and containment of epidemics and also aid in the prevention of cyberterrorism, identity theft and a host of other web based crimes and societal threats. For example, for the past few years, scientist have been able to pinpoint flu outbreaks by simply tracking the amount of people who goolgled “flu”.

In a study of social media sharing, University of Pennsylvania psychologist Andrew Schwartz, has found that most posts—especially by women—contain emotional language.  Schwartz and his team study social media, posts and updates in an effort to learn about people.  By studying the information shared via these platforms, he believes a general understanding of the physical and mental wellbeing of individuals and communities can be gained.  Schwartz further explains that by analyzing an individual’s Facebook profile or Twitter feed he can accurately predict intimate traits about that person that are not being revealed intentionally.

Canadian and French researchers  have partnered and are attempting to develop an algorithm that can catch warning signs and possibly predict dangerous behaviors. University of Ottawa’s Diana Inkpen has developed a project called “Social Web Mining and Sentiment Analysis For Mental Illness Detection.”  Essentially, what Inkpen and her team are attempting to do is develop an algorithm that screens online posts for warning signs of mental illness. It is the goal of this team, to create a set of tools that doctors, psychologists, counselors and mental health professionals can use to identify concerning or dangerous patterns by using social media posts.

This type of information gathering—while It may seem creepy and horribly invasive is proving to be pivotal in assisting to make our world a slightly safer place.