Exploring the Potential of Social Media Data in Interpreting Traffic Congestion: A Case Study of Jiangsu Freeways

Social network platforms provide more open data interfaces for the public, which could be a source for various applications, especially for mining traffic jams. This study aims to identify the congestion-prone areas of Jiangsu freeways via Sina Weibo and further analyze the characteristics of the areas. Specifically, this study first obtained congestion-related microblogs from Sina Weibo using a web crawler, then a series of data preprocessing methods were applied to optimize the information. Second, the document frequency-based method was adopted to figure out the congestion-prone areas. Afterward, the characteristics of the areas were explored by the Apriori-based algorithm and the co-occurrence network diagram. Finally, the experiment results were testified with the ground truth from the authority “Jiangsu Freeway 96777.” It indicated that the identified congestion-prone areas fit well with those from the ground truth. Therefore, the proposed methods provide an auxiliary way to interpret traffic congestion for traffic operators.