Taxi Global Positioning System Data in Urban Road Network: A Methodology to Identify Key Road Clusters Based on Travel Speed–Traffic Volume Correlation

Abstract

Urban traffic congestion is one of the urban diseases that needs to be solved urgently. Research has already found that a few road segments can significantly influence the overall operation of the road network. Traditional congestion mitigation strategies mainly focus on the topological structure and the transport performance of each single key road segment. However, the propagation characteristics of congestion indicate that the interaction between road segments and the correlation between travel speed and traffic volume should also be considered. The definition is proposed for “key road cluster” as a group of road segments with strong correlation and spatial compactness. A methodology is proposed to identify key road clusters in the network and understand the operating characteristics of key road clusters. Considering the correlation between travel speed and traffic volume, a unidirectional-weighted correlation network is constructed. The community detection algorithm is applied to partition road segments into key road clusters. Three indexes are used to evaluate and describe the characteristic of these road clusters, including sensitivity, importance, and IS. A case study is carried out using taxi GPS data of Shanghai, China, from May 1 to 17, 2019. A total of 44 key road clusters are identified in the road network. According to their spatial distribution patterns, these key road clusters can be classified into three types—along with network skeletons, around transportation hubs, and near bridges. The methodology unveils the mechanism of congestion formation and propagation, which can offer significant support for traffic management.

TransBigData: A Python package for transportation spatio-temporal big data processing, analysis and visualization

“…Further, a variety of processing methods for each stage of transportation spatio-temporal big data analysis are included in the code base. TransBigData provides clean, efficient, flexible, and easy to use API, allowing complex data tasks to be achieved with concise code, and it has already been used in a number of scientific publications (Li et al, 2021;Yu, Zhang, Li, Sui, et al, 2020;Yu et al, 2021).…”