Dynamic characteristics of tweeting and tweet topics

Twitter, having more than 200 million world users and more than 4 million Korean users, is still growing fast. Because Twitter users can ‘tweet’ about any topic within the 140-character limit, and other users who follow the users and see the tweets can ‘retweet’ them, Twitter is regarded as a new medium of transferring and sharing information. Nevertheless, the propensities of Twitter users to tweet or to retweet still remain unclear. In order to investigate these propensities, we propose a simple model for the dynamics of the total number of tweets about specific topics. We then observe that the topics can be categorized into three kinds according to predictability and sustainability: predictable events, unpredictable events, and sustainable events. Comparing model results with real data, we infer the tweet propensities motivated by external causes as well as retweet propensities.