RT^2M: Real-Time Twitter Trend Mining System

Abstract

The advent of social media is changing the existing information behavior by letting users access to real-time online information channels without the constraints of time and space. It also generates a huge amount of data worth discovering novel knowledge. Social media, therefore, has created an enormous challenge for scientists trying to keep pace with developments in their field. Most of the previous studies have adopted broad-brush approaches which tend to result in providing limited analysis. To handle these problems properly, we introduce our real-time Twitter trend mining system, RT2M, which operates in real-time to process big stream datasets available on Twitter. The system offers the functions of term co-occurrence retrieval, visualization of Twitter users by query, similarity calculation between two users, Topic Modeling to keep track of changes of topical trend, and analysis on mention-based user networks. We also demonstrate an empirical study on 2012 Korean presidential election. The case study reveals Twitter could be a useful source to detect and predict the advent and changes of social issues, and analysis of mention-based user networks could show different aspects of user behaviors.