You're happy, I'm happy: diffusion of mood expression on twitter

We study the propagation of expression of moods from one individual to another in social media. Specifically, we utilize an ensemble of more than 200 moods, and follow their trails of retweets on Twitter. We examine the diffusion of these moods through the characteristics of the author of the retweeted post, of the post, and its linguistic style and topical content. We observe that moods of high valence and low to moderate activation propagate the most. Our findings also indicate that mood-bearing posts from individuals of medium #followers show highest diffusion, instead of elite users. Further, moods in postings with high self-attentional focus, posted by socially interactive women, and with fewer links diffuse the most. Finally, we leverage these characteristics to build a linear regression model that can predict extent of diffusion in a mood-bearing post. The model yields accuracy of 75% showing that user's attributes as well as post's language characteristics are key factors driving the diffusion of moods in social media.