Traffic event detection framework using social media

Traffic incidents are one of the leading causes of non-recurrent traffic congestions. By detecting these incidents on time, traffic management agencies can activate strategies to ease congestion and travelers can plan their trip by taking into consideration these factors. In recent years, there has been an increasing interest in Twitter because of the real-time nature of its data. Twitter has been used as a way of predicting revenues, accidents, natural disasters, and traffic. This paper proposes a framework for the real-time detection of traffic events using Twitter data. The methodology consists of a text classification algorithm to identify traffic related tweets. These traffic messages are then geolocated and further classified into positive, negative, or neutral class using sentiment analysis. In addition, stress and relaxation strength detection is performed, with the purpose of further analyzing user emotions within the tweet. Future work will be carried out to implement the proposed framework in the West Midlands area, United Kingdom.