Sensing and detecting traffic events using geosocial media data: A review

Abstract Social media platforms, or social networks, have allowed millions of users to post online content about topics related to our daily lives. Traffic is one of the many topics for which users generate content. People tend to post traffic related messages through the ever-expanding geosocial media platforms. Monitoring and analyzing this rich and continuous user-generated content can yield unprecedentedly valuable traffic related information, which can be mined to extract traffic events to enable users and organizations to acquire actionable knowledge. A great number of literature has reported on the methods developed for detecting traffic information from social media data, especially geosocial media data when geo-tagged. However, a systematic review to synthesize the state-of-the-art developments is missing. This paper presents a systematic review of a wide variety of techniques applied in detecting traffic events from geosocial media data, arranged based on their adoption in each stage of an event detection framework developed from the literature review. The paper also highlights some challenges and potential solutions. The aim of the paper is to provide a structured view on current state-of-art of the geosocial media based traffic event detection techniques, which can help researchers carry out further research in this area.