Topic Tracker : Shape-based Visualization for Trend and Sentiment Tracking in Twitter

Abstract

In recent years there has been a continuous development of social media services on the web. Unprecedented success and active usage of these services result in massive amounts of user-generated data. Visual representation of these large amounts of unevenly distributed time series data is a challenging task, especially while preserving access to individual data points. Our hypothesis is that shape-based visual representations have advantages over established time series compression visualizations like Two-Tone Pseudo Coloring or line graphs. In this paper we present a shape-based visualization for trend and sentiment tracking of user-defined topics in the Twitter data stream. We use glyphs to visualize the appearance and sentiment of tweets on a timeline and enable analysts to keep track of the trend of their defined topic and the corresponding sentiment expressed by the Twitter users.