While Instagram data can already be used to guess your age, a new research paper shows how it might also be used to check upon your mental health. Using a set of machine learning tools and several dozen users' Instagram feeds, a team of researchers from Harvard and the University of Vermont have built a model that can accurately spot signs of clinical depression. By reviewing "color analysis, metadata components, and algorithmic face detection," in each user's feed, the model was able to correctly identify which Instagrammers showed symptoms of depression about 70 percent of the time, even before they had been clinically diagnosed.
The model had to sift through 43,950 photos from 166 different users in order to make its predictions. And, before everyone becomes an amateur Instagram psychologist, the research team notes that their model isn't meant to be a definitive diagnosis of depression just yet. Instead, the paper notes that the model could be used for "early screening and detection of mental illness" and could one day "serve as a blueprint for effective mental health screening in an increasingly digitalized society." In other words: if your phone's digital assistant has access to your Instagram feed, it might one day be able to tell if you've been seeming blue lately.
And that "blue" could be in the literal sense -- although the model took many factors into account, the study found that depressed individuals tended to gravitate towards the the blue-grey or black-and-white filters like Crema or Inkwell, while healthy folks preferred filters with warm, bright tones.