Depression, Economics, and Population in Social Media

• Activity Number: 179
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Defense and National Security
• Author(s): Elizabeth Hohman*+ and David Marchette and Glen Coppersmith
• Companies: Naval Surface Warfare Center and Naval Surface Warfare Center and Johns Hopkins University
• Keywords: text analysis ; social media analysis ; Twitter

Abstract

Given a set of Twitter users who were diagnosed with one of several types of depression, we look at how their posts differ from a sample of the general population of Twitter users. The types of depression considered include Seasonal Affective Disorder (SAD) and Post Traumatic Stress Disorder (PTSD). We look at differences in the probabilities of word usage in these populations. Additionally, we use selected posts to build a depression classifier which we then run on all geo-tagged tweets in the continental United States. This results in county-wide estimates for the percent of depressed users. We look at these estimates along with economic and social factors as well as seasonal trends. Results are presented along with model caveats.