A study published in the EPJ Data Science Journal revealed that photos on Instagram, which are bluer or darker in color or even those which are less saturated might reveal if a person is struggling with mental health issues.
Instagram photos reveal predictive markers of depression #socialmedia #machinelearning https://t.co/v3png1j61N pic.twitter.com/1J433q94YL
— EPJ Data Science (@epj_ds) August 8, 2017
The study authors, Professor Chris Danforth from the University of Vermont and Dr. Andrew G. Reece, who works as a machine learning engineer in Silicon Valley, have examined Instagram data from 166 individuals and uncovered what the successful markers are for depression.
"Statistical features were computationally extracted from 43,950 participant Instagram photos, using color analysis, metadata components, and algorithmic face detection," an extract from the study reads.
The researchers were then able to predict which photos showed signs of depression using the study mode.
They did not take into consideration captions or comments, however, by examining the pixel of each picture, they were able to find out that pictures of a darker, bluer or grayer color were signs of depression.
The methodology used meant that people with depression were correctly identified 70 percent of the time.
According to the study, this is better than the average doctor who correctly diagnoses depression 50 percent of the time.
"We were looking to identify what behaviors people are exhibiting potentially without them even realizing," Professor Danforth, said.
The computer model found that depressed users of Instagram would post bluer, darker and grayer pictures, which were also unfiltered.
They tended to have fewer likes, and when a filter was used, the preference was a black-and-white color.
Non-depressed people who use Instagram tended to post pictures which were warm and sunny.
"It's a proof of concept and for the particular individuals we studied, this set of predictors works for them. Whether or not it would translate to average person's Instagram feed, we don't know," Professor Danforth said.
There have been other studies which have examined whether social media can display whether the user is in fact depressed.
One study was able to predict whether mothers were at risk of postpartum depression based on their social media posts.