Analysis of Emotions in Real-time Twitter Streams

Abstract

 The purpose of this paper is to present EmoTwitter 2.0, a system for visualizing discussions and emotions of Twitter users in real time over a specific geographical location. The system, given location information as input, streams users’ tweets posted in real time using the Twitter API. In addition, using content analysis, it extracts and visualizes the most frequent words and emotional content of the streamed tweets. Through demonstrations of potential use cases and testing, the system has shown to have practical applicability. It provides an opportunity to easily visualize and compare the discussions of people on Twitter in certain geographical locations, which can be useful, for instance, in targeted messaging.