Traffic zone division using mobile billing data

Traffic zoning which could simplify complex urban traffic network is a fundamental work in urban planning and transportation management. Traditional traffic zone division is mainly on census data which cost tremendous resources but could only cover part of people and areas in city. This paper develops a comprehensive approach of traffic zone division on mobile billing data which owns the advantages of (1). high coverage, (2). cost effective and (3). up to date. Land use information obtained from phone call volume and commuting volume as well as spatial obstacles represented as Voronoi distance are taken into account in similarity measurements. Clustering measurements and traffic zone factors are employed to evaluate our approach. Experiments show good performance of our method in both dealing with spatial obstacles and traffic zone measurements.