A unified spatial multigraph analysis for public transport performance

Yaoli Wang,Yu Liu,Di Zhu,Ganmin Yin,Zhou Huang

doi:10.1038/s41598-020-65175-x

Yaoli Wang, Yu Liu + Show 3 more

Open Access

https://doi.org/10.1038/s41598-020-65175-x

Copy DOI

Abstract

Public transport performance not only directly depicts the convenience of a city’s public transport, but also indirectly reflects urban dwellers’ life quality and urban attractiveness. Understanding why some regions are easier to get around by public transport helps to build a green transport friendly city. This paper initiates a new perspective and method to investigate how public transport network’s morphology contributes significantly to its performance. We investigate the significance of morphology from the perspective of graph classification – by extracting the typical local structures, called “motifs”, from the multi-modal public transport multigraph. A motif is seen as a certain connectivity pattern of different transport modes at a local scale. Combinations of various motifs decide the output of graph classification, particularly, public transport performance. We invent an innovative method to extract motifs on complex spatial multigraphs. The proposed method is adaptable to unify complex factors into one simple form for swift coding, and depends less on observational data to handle data unavailability. In the study area of Beijing, we validate that the measure can infer varied public transport efficiencies and access equalities of different regions. Some typical areas with undeveloped or unevenly distributed public transport are further discussed.

Highlights

Public transport performance directly depicts the convenience of a city’s public transport, and indirectly reflects urban dwellers’ life quality and urban attractiveness
We hypothesize that morphology of the public transport (PT) network contributes significantly to its performance heterogeneity and investigate the association from a new perspective – graph classification
There is a good justification why morphology should be studied: 1) morphology is an effective factor as both other studies have discovered[4,5,6,7,8,9] and this paper demonstrates below, 2) it avoids an extensive amount of data and data unavailability, 3) it provides the foundation for a unified and swift coding as this study proposes

Summary

Introduction

Public transport performance directly depicts the convenience of a city’s public transport, and indirectly reflects urban dwellers’ life quality and urban attractiveness. This paper initiates a new perspective and method to investigate how public transport network’s morphology contributes significantly to its performance. A significant contribution of this work is to extract motifs from PT multigraphs for its performance assessment, which provides a swift and unified encoding of heterogeneous network properties. This inherits the concept of graph classification, a process to predict the category of a graph by the combination of motifs – typical functional structures that either occur significantly more than random[17], or function-wise contribute to the identification of a graph[18]. Urban planners can apply this method to facilitate the pre-assessment of PT network performance by investigating the extracted motifs

Objectives

Methods

Results