Abstract

In this study we investigate the relationships between a variety of automatically inferred latent user properties, opinions and emotions on a large sample of users in online social network. We first correlate user demographics and personality with the emotional profile emanating from user tweets. We then analyse the relationships between predicted user properties and user-environment emotional contrast estimated over various neighbourhoods including friends, retweeted and mentioned users. We demonstrate that many perceived user demographic traits correlate with the emotional contrast between users and their neighbours. We show that users predicted to be single or those with lower income tend to amplify sad tweets, but older users or those with children dampen them; users satisfied with life dampen angry tweets whereas users dissatisfied with life amplify them; optimists amplify joyful tweets while pessimists dampen them. Based on these correlations, we show that it s possible to accurately predict a range of perceived demographic traits for a given user, based solely on the emotional tone emanating from that user and his neighbours.