Detection of potential traffic jam based on traffic characteristic data analysis

Traffic jam is one of the big and complex problems which happens in many big cities around the world. The advances in technology have enabled tracking moving objects such as vehicles move on road networks. Those data is called spatio-temporal data. By using data mining techniques, analysis on traffic characteristics can be conducted to detect potential area of traffic jam. This information is particularly useful for transportation bureau to make appropriate decision regarding traffic policy. In this paper, we propose a method to analyze data of traffic characteristics which consists of clustering traffic characteristic data, ranking process, and analyzing the cluster with respect to the traffic jam criteria. A visualization is also applied to give better understanding of the analysis results. The experiments show that data mining of spatio-temporal data can be used to analyze traffic characteristic, yet it could neither identify the source of the traffic jam nor the congested routes.