Trend Visualization on Twitter: What's Hot and What's Not?

摘要

Twitter is a social networking service in which users can create short messages related to a wide variety of subjects. Certain subjects are highlighted by Twitter as the most popular subjects and are known as trending topics. In this paper, we study the visual representation of these trending topics to maximize the information toward the users in the most effective way. For this purpose, we present a new visual representation of the trending topics based on dynamic squarified treemaps. In order to use this visual representation, one needs to determine (preferably forecast) the speed at which tweets on a particular subject are posted and one needs to detect acceleration. Moreover, one needs efficient ways to relate topics to each other when necessary, so that clusters of related trending topics are formed to be more informative about a particular subject. We will outline the methodologies for determining the speed and acceleration, and for clustering. We show that the visualization using dynamic squarified treemaps has many benefits over other visualization techniques.