Authors:  Peter Dodds and Chris Danforth, Department of Mathematics & Statistics

Abstract:  The importance of quantifying the nature and intensity of emotional states at the level of populations is evident: we would like to know how individuals feel so that we may improve public policy, build more successful organizations, and more fully understand economic and social phenomena. Using human evaluation of the happiness of words, we analyze a diverse set of texts which reflect human experience including millions of weblogs and billions of status updates from Twitter. Among numerous observations we will present, we find that expressed happiness rises and falls with age and distance from the Earth's equator; the 2008 Presidential Election was the happiest day in the blogosphere in the last 4 years; Twitter users' status updates increase in happiness in the manner of a phase transition as a function of the number of others listed as following them; and written communication exhibits a pro-social nature.