Data-driven interpretation on interactive and nonlinear effects of the correlated built environment on shared mobility
Journal article
Gao, Kun, Yang, Ying, Gil, Jorge and Qu, Xiaobo. (2023). Data-driven interpretation on interactive and nonlinear effects of the correlated built environment on shared mobility. Journal of Transport Geography. 110, p. Article 103604. https://doi.org/10.1016/j.jtrangeo.2023.103604
Authors | Gao, Kun, Yang, Ying, Gil, Jorge and Qu, Xiaobo |
---|---|
Abstract | Understanding the usage demand of shared mobility systems in different areas of a city and its determinants is crucial for planning, operation and management of the systems. This study leverages an unbiased data-driven approach called accumulated effect analysis for examining the complex (nonlinear and interactive) effects of correlated built environment factors on the usage of shared mobility. Special research emphasis is given to unraveling the complex effects using an unbiased and data-driven approach that can overcome the impacts of correlations among built environment factors. Based on empirical analysis of synthetic data and a field dataset about dockless bike sharing systems (DLBS), results demonstrate that the method of partial dependency analysis prevalent in the relevant literature, will result in biases when investigating the effects of correlated built environment factors. In comparison, accumulated local effect analysis can appropriately interpret the effects of correlated built environment factors. The main effects of many built environment factors on the usage of DLBS present nonlinear and threshold patterns, quantitively revealed by accumulated local analysis. The approach can reveal complex interaction effects between different built environment factors (e.g., commercial service and education facility, and metro station coverage and living facility) on the usage of DLBS as well. The interactions among two built environment factors could even change with the values of the factors rather than invariant. The outcomes offer a new approach for revealing complex influences of different built environment factors with correlations as well as in-depth empirical understandings regarding the usage of DLBS. |
Keywords | interpretable machine learning; built environment factors; shared mobility systems; interactive effects |
Year | 2023 |
Journal | Journal of Transport Geography |
Journal citation | 110, p. Article 103604 |
Publisher | Elsevier Ltd |
ISSN | 0966-6923 |
Digital Object Identifier (DOI) | https://doi.org/10.1016/j.jtrangeo.2023.103604 |
Scopus EID | 2-s2.0-85159278051 |
Open access | Published as ‘gold’ (paid) open access |
Page range | 1-14 |
Publisher's version | License File Access Level Open |
Output status | Published |
Publication dates | |
Online | 23 May 2023 |
Publication process dates | |
Accepted | 07 May 2023 |
Deposited | 01 Apr 2025 |
Additional information | © 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). |
https://acuresearchbank.acu.edu.au/item/9186y/data-driven-interpretation-on-interactive-and-nonlinear-effects-of-the-correlated-built-environment-on-shared-mobility
Download files
Publisher's version
OA_Gao_2023_Data_driven_interpretation_on_interactive_and.pdf | |
License: CC BY 4.0 | |
File access level: Open |
1
total views0
total downloads1
views this month0
downloads this month