Emotion Expression on Social Networking Sites: Exploring Mood Profiles and Depression

Depression can be detected from the language people use on social media. This thesis explored patterns in the way people express emotion online and how emotion patterns can be used to identify depression from status updates. Language is complex, and the emotion expressed in status updates did not clearly reflect experienced emotion at a daily level. Emotion patterns over time were more informative. For Facebook users, extreme fluctuations in the amount of negative emotion words between consecutive status updates was predictive of depression. However, on Twitter, using a broad range of negative emotion words was protective for mental health.