Word occurrences and emotions in social media: Case study on a Twitter corpus
Ivan Đunđer, Marina Horvat, Sergej Lugović · 2016
Twitter is currently the most popular tool for social interaction and real-time information exchange. Outreach and importance of individual accounts is measured by the number of their followers. The aim of this paper is to investigate the applicability and usefulness of corpora containing textual and visual information for the purpose of machine observation of Twitter activities. The results in the presented analytic research are based on a data set of more than 16000 tweets collected from 22 startup founders' Twitter accounts with a large number of followers over a four-month period. Word usage in tweets was examined with natural language processing (NLP) techniques, applying word occurrence analyses and a manual qualitative evaluation of frequent words within the data set, primarily focusing on the distribution of words. Furthermore, profile pictures of Twitter accounts were collected in order to conduct a facial emotion analysis and emotion mining. Estimated basic emotional states were statistically compared with the number of tweets posted and the number of new followers gained during the observed timespan.