My Ever Changing Moods: Sentiment-based Event Detection on the Cloud

Twitter is a globally used micro-blogging platform with hundreds of millions of tweets sent every day. Many researchers have explored Twitter analytics across a wide range of areas such as topic modeling, sentiment analysis, event detection, as well as the application of Twitter for a variety of domain-specific application areas, e.g. disaster management. One area that has not been explored is how changes in sentiment can be used to identify events. In this paper we present a scalable Cloud-based platform for harvesting, processing, analyzing and visualizing large-scale Twitter data. We focus especially on how changes in sentiment can be used to identify events in given contexts. What is novel is that the events that are detected are not dependent explicitly on the topic of any given tweet, but entirely on the change in sentiment. This offers new capabilities for event detection that have hitherto not been explored. To illustrate the approach, we present case studies related to sporting events identified entirely through changing sentiment with specific focus on the 2014 FIFA World Cup of Soccer and the 2015 World Cup of Cricket.