Abstract
Real-time analysis of traffic data is a key challenge in intelligent transportation system. It aims at discovering useful traffic patterns that can help decision makers better manage the transportation system and test and introduce new policies. Discovered patterns can also be used to support road users to reach their destination safely and with reasonable commuting time. In this paper, a number of key challenges associated with transportation systems and possible solutions are discussed. A method that analyzes real-time traffic data to predict future status of traffic flow and incidents is introduced. The proposed method includes three phases: offline, real-time, and decision support phases. In this paper, a decision tree classification model is constructed and validated for an accident dataset. Possible benefits of using the constructed model are demonstrated using results of the classification analysis.