Applying and Evaluating Data-Driven Fine Grid Partitioning Methods for Traffic Analysis Zones

Abstract

Fine grid management is one of the important development directions of transportation planning that has not been fully considered in previous literature. This paper explores the application of the fine grid management method in transportation planning. Based on a case study of Chuanhui, China, this paper proposes a data-driven fine grid partitioning method for determining traffic analysis zones (TAZs). The TAZs are partitioned based on quadrilateral and hexagonal grids. This paper also summarizes a set of criteria for evaluating the impact of different fine grid partitioning methods based on the geographically and temporally weighted regression (GTWR) model. The results show that our fine grid partitioning method for determining TAZs based on quadrilateral grids can achieve a relatively low level of predicted value bias and variable correlation degree bias when the number of TAZs is larger, and it has obvious advantages. Finally, policy implications are proposed to promote the refinement of transportation planning.