What's Happening: Finding Spontaneous User Clusters Nearby Using Twitter

Abstract

Twitter is a public social service that allows users to share information as short-text messages. Previous researchers have tried to analyze the information available on Twitter to discover topic trending. However, these topics are associated with the whole network, and are not associated to a particular place. In this work, we propose a scheme to discover places where something is happening by analyzing geo-tagged tweets in a timely manner. By using the proposed scheme, users will be able to recognize spontaneous events happening nearby, and be notified about it through their smart phones. Experimental results show that by using geo-tagged information associated to tweets, events related to a particular place can be detected using clustering techniques and semantic interpretation of keywords.