Nature and causes of urban traffic congestion - a case study of London

This paper presents an empirical approach to investigate the nature and causes of urban traffic congestion. Compared with freeways, urban network traffic is relatively less studied due to its complexity and availability of required data. This paper introduces the data resources the authors have and the research they have been conducting to understand the characteristic of urban traffic in Greater London Area, UK. Our traffic data include journey time estimates from Automatic Number Plate Recognition (ANPR) system and GPS (Global Positioning System) equipped vehicles; traffic counts, incident records from London Traffic Information System (LTIS), and weather information from UK MET office. With the traffic data, the authors introduce the use of visualization to retrieve the spatio-temporal feature of urban traffic. We also present the use of linear regression model to diagnose observed congestion and attribute them to different causes. In particular, they distinguish the observed congestion into two main components: one due to recurrent factors and the other due to non-recurrent factors. The methodologies are illustrated through a case study of Greater London Area. It is found that about 15% of the observed congestion in the region is due to non-recurrent factors such as accidents, unplanned roadwork, special events, and strikes. The study presented herein will be valuable for transport policy evaluation and appraisal.