2023
Lyu, Tao; Wang, Yuanqing; Ji, Shujuan; Feng, Tao; Wu, Zhouhao
A multiscale spatial analysis of taxi ridership Journal Article
In: Journal of Transport Geography, vol. 113, pp. 103718, 2023, ISSN: 0966-6923.
Abstract | Links | BibTeX | Tags: Influencing factors, Multiscale geographically weighted regression, Scale effect, spatial heterogeneity, Taxi ridership
@article{LYU2023103718,
title = {A multiscale spatial analysis of taxi ridership},
author = {Tao Lyu and Yuanqing Wang and Shujuan Ji and Tao Feng and Zhouhao Wu},
url = {https://www.sciencedirect.com/science/article/pii/S0966692323001904},
doi = {https://doi.org/10.1016/j.jtrangeo.2023.103718},
issn = {0966-6923},
year = {2023},
date = {2023-01-01},
journal = {Journal of Transport Geography},
volume = {113},
pages = {103718},
abstract = {Taxi plays a supplement role in sustainable development of urban public transport systems. However, the extent to which the built environment affects taxi ridership at various spatial scales deserves further exploration because understanding the true spatial heterogeneity across a varying scale could be valuable for both global and localized policy decision-makings. In this study, we attempt to analyze and discuss the spatial predictors of taxi ridership by utilizing an multiscale geographically weighted regression (MGWR) model and comparing the model's performance to that of ordinary least square (OLS) and geographically weighted regression (GWR) models. Using the taxi data of Xi'an city, we found that the MGWR model could explain 81.8% of the total taxi ridership fluctuations and allows localized and targeted policy makings to help taxi drivers search for passengers and to improve passengers' taxi-hailing experiences in specific districts.},
keywords = {Influencing factors, Multiscale geographically weighted regression, Scale effect, spatial heterogeneity, Taxi ridership},
pubstate = {published},
tppubtype = {article}
}
Taxi plays a supplement role in sustainable development of urban public transport systems. However, the extent to which the built environment affects taxi ridership at various spatial scales deserves further exploration because understanding the true spatial heterogeneity across a varying scale could be valuable for both global and localized policy decision-makings. In this study, we attempt to analyze and discuss the spatial predictors of taxi ridership by utilizing an multiscale geographically weighted regression (MGWR) model and comparing the model’s performance to that of ordinary least square (OLS) and geographically weighted regression (GWR) models. Using the taxi data of Xi’an city, we found that the MGWR model could explain 81.8% of the total taxi ridership fluctuations and allows localized and targeted policy makings to help taxi drivers search for passengers and to improve passengers’ taxi-hailing experiences in specific districts.
2019
Gan, Z.; Feng, T.; Yang, M.; Timmermans, H.; Luo, J.
Analysis of Metro Station Ridership Considering Spatial Heterogeneity Journal Article
In: Chinese Geographical Science, vol. 29, iss. 6, 2019, ISSN: 1993064X.
Abstract | Links | BibTeX | Tags: Built environment, rapid transit ridership, spatial heterogeneity, spatial models, station level
@article{Gan2019,
title = {Analysis of Metro Station Ridership Considering Spatial Heterogeneity},
author = {Z. Gan and T. Feng and M. Yang and H. Timmermans and J. Luo},
doi = {10.1007/s11769-019-1065-8},
issn = {1993064X},
year = {2019},
date = {2019-01-01},
journal = {Chinese Geographical Science},
volume = {29},
issue = {6},
abstract = {This study aims to explore the role of spatial heterogeneity in ridership analysis and examine the relationship between built environment, station attributes and urban rapid transit ridership at the station level. Although spatial heterogeneity has been widely acknowledged in spatial data analysis, it has been rarely considered in travel behavior studies. Four models (three global models-ordinary least squares (OLS), spatial lag model (SLM), spatial error model (SEM) and one local model-geographically weighted regression (GWR) model) are estimated separately to explore the relationship between various independent variables and station ridership, and identify the influence of spatial heterogeneity. Using the data of built environment and station characteristics, the results of diagnostic identify evidence the existence of spatial heterogeneity in station ridership for the metro network in Nanjing, China. Results of comparing the various goodness-of-fit indicators show that, the GWR model yields the best fit of the data, performance followed by the SEM, SLM and OLS model. The results also demonstrate that population, number of lines, number of feeder buses, number of exits, road density and proportion residential area have a significant impact on station ridership. Moreover, the study pays special attention to the spatial variation in the coefficients of the independent variables and their statistical significance. It underlines the importance of taking spatial heterogeneity into account in the station ridership analysis and the decision-making in urban planning.},
keywords = {Built environment, rapid transit ridership, spatial heterogeneity, spatial models, station level},
pubstate = {published},
tppubtype = {article}
}
This study aims to explore the role of spatial heterogeneity in ridership analysis and examine the relationship between built environment, station attributes and urban rapid transit ridership at the station level. Although spatial heterogeneity has been widely acknowledged in spatial data analysis, it has been rarely considered in travel behavior studies. Four models (three global models-ordinary least squares (OLS), spatial lag model (SLM), spatial error model (SEM) and one local model-geographically weighted regression (GWR) model) are estimated separately to explore the relationship between various independent variables and station ridership, and identify the influence of spatial heterogeneity. Using the data of built environment and station characteristics, the results of diagnostic identify evidence the existence of spatial heterogeneity in station ridership for the metro network in Nanjing, China. Results of comparing the various goodness-of-fit indicators show that, the GWR model yields the best fit of the data, performance followed by the SEM, SLM and OLS model. The results also demonstrate that population, number of lines, number of feeder buses, number of exits, road density and proportion residential area have a significant impact on station ridership. Moreover, the study pays special attention to the spatial variation in the coefficients of the independent variables and their statistical significance. It underlines the importance of taking spatial heterogeneity into account in the station ridership analysis and the decision-making in urban planning.