During my postdoctoral position at Northwell Health, I worked with Dr Michael L. Birnbaum on analyzing Instagram archives provided by patients with schizophrenia spectrum disorder (SSD) and healthy volunteers. These archives contained rich data on each participants Instagram usage, e.g., a complete history of posts, private messages, and interactions with other users’ content (e.g., likes and comments).

We focused on analyzing image characteristics in particular due to the nature of Instagram: it is a social media platform for sharing photo and video content and provides easy tools for manipulating these to a user’s liking, e.g., apply filters. We used OpenCV to extract image features around color, aspect ratio, and human faces depicted in the images. We further considered basic information of each user, e.g., followers.

We found that individuals who are diagnosed with SSD, post less saturated and less colorful images on Instagram. Further, their profiles showed a lower ratio of followers to following compared to HV participants.

Publication: Katrin Hänsel, Inna Wanyin Lin, Michael Sobolev, Whitney Muscat, Sabrina Yum-Chan, Munmun De Choudhury, John M. Kane, Michael L. Birnbaum. Utilizing Instagram Data to Identify Usage Patterns Associated with Schizophrenia Spectrum Disorders In: Frontiers in Psychiatry 12, 2021 (paper)

Relevant Skills:
Image features with OpenCV, social media features, inferential statistics