Elsevier

Cities

Volume 101, June 2020, 102689
Cities

Space-time dynamics of cab drivers' stay behaviors and their relationships with built environment characteristics

https://doi.org/10.1016/j.cities.2020.102689Get rights and content

Highlights

  • A spatiotemporal analysis of cab drivers' stay behaviors using taxi GPS data

  • The collective stay behavior across urban location follows an exponential decay

  • The stay frequency on weekdays and on weekends shows similar spatial patterns

  • The built environment can well explain the spatial variations of stay behavior

Abstract

Understanding cab drivers' stay activities is essential for planning and managing certain urban facilities. This study analyzes cab drivers' stay behaviors using a taxi GPS trajectory dataset collected in Wuhan, China. By extracting cab drivers' stay activities from the dataset, we measure the activity frequency at the level of traffic analysis zones (TAZs) and examine their spatiotemporal dynamics. We then derive several built environment indicators and assess their associations with these activities using ordinary least squares regression (OLS) and geographically weighted regression (GWR) models. According to the results, the stay frequency decays dramatically over the TAZs, indicating that these activities tend to be concentrated in particular areas of the city. The rates of decay, as reflected by the rank-size and power-law distributions, are similar on weekdays and weekends. Cab drivers' stay activities exhibit similar spatial patterns during the same period on weekdays and weekends. The adjusted R-squared of OLS is 0.742 for weekdays and 0.676 for weekends, which suggests a close relationship between stay activities and built environment characteristics. The GWR models further reveal the spatial variations of the activity-environment linkage across the study area. The study provides useful insights that support future urban design and transport planning.

Introduction

As one of the key components of urban transportation systems, taxis provide an all-weather, convenient, and personalized travel service for urban residents. Improving the efficiency of taxi services based on the knowledge of cab drivers' behavioral patterns is essential for transport planning and management of urban facilities (Kang & Qin, 2016). When and where taxis tend to stay are important but often understudied aspects of drivers' behaviors. It is necessary to investigate cab drivers' stay behaviors for several reasons. First, cab drivers' stay behaviors reflect specific types of activities that are linked to the spatial configuration of urban facilities. For instance, many vacant taxis queue up in specific areas for purposes including refueling or dining. Improving our knowledge of these behaviors could help cities better satisfy the needs of taxi drivers. Second, cab drivers' stay behaviors also reflect their operation behaviors (e.g., waiting for passengers), which in turn convey useful information about taxi supply and the areas where potential demand might be high.

Previous studies on cab drivers' activities mainly rely on data collected through travel surveys, questionnaires or interviews. Most of these studies focus on analyzing driving behavior without paying enough attention to drivers' stay activities (Kalhori, Foroughinia, & Ziapour, 2017; Ma, Yan, Huang, & Abdel-Aty, 2010; Newnam, Mamo, & Tulu, 2014; Shi, Tao, Li, Xiao, & Atchley, 2014; Zhang et al., 2018). The few studies that analyze cab drivers' dining behavior are based on questionnaire and survey data (Song et al., 2012; Zhang, Zheng, Lu, & Chai, 2009). This has led to a scarcity of knowledge on cab drivers' stay behavior, an important dimension of their activities beyond movements along road networks.

Overall, studies on cab drivers' stay behavior are sparse. As data collection of surveys and questionnaires are costly and time consuming, they usually cover small sample sizes, which may not be able to delineate cab drivers' behavior at a large scale. In recent years, the explosive growth of GPS-based taxi trajectories has provided massive opportunities for analyzing cab drivers' behaviors. Compared to traditional data based on travel surveys or questionnaires, GPS-based taxi trajectory data are collected by tens of thousands of taxis simultaneously at low costs and high efficiency. A strand of studies on quantifying cab driver's mobility patterns based on GPS taxi trajectory data is continuously emerging. For instance, cab drivers' operational behavior (Gao, Jiang, & Xu, 2018; Liu, Andris, & Ratti, 2010; Manley, Addison, & Cheng, 2015; Tang et al., 2016), refueling behavior (Niu, Liu, Fu, Liu, & Lang, 2016; Zhang, Yuan, Wilkie, Zheng, & Xie, 2015), and shift handover behavior (Sun & Yu, 2014; Zhang et al., 2015) have been investigated based on GPS taxi trajectory data. Despite the various types of cab drivers' behaviors examined in these studies, little effort has been devoted to systematically investigating cab drivers' stay behaviors.

In this paper, we use GPS-based taxi trajectory data collected in Wuhan, China to investigate three aspects of cab drivers' stay behaviors, namely, the regularities of cab drivers' stay behaviors across space, the spatiotemporal variations of stay frequency, and the relationship between stay frequency and certain built environment characteristics. Our specific research questions are as follows: (1) What are the key characteristics of the spatial distributions of these stay activities? (2) How do the spatial patterns of such activities change over time? and (3) How are cab drivers' collective stay patterns related to the underlying built environment? To answer these questions, we apply a trajectory prepossessing method to identify taxis' stay points, from which the intensity and evolution of cab drivers' stay activities are mapped and analyzed. Two models are utilized to measure stay activity frequency in this study, namely the rank-size distribution and the frequency probability distribution. Second, the spatial distributions of stay activities during different periods are analyzed. Finally, we apply ordinary least squares regression (OLS) and geographically weighted regression (GWR) models to investigate the relationships between stay activities and certain built environment characteristics.

Section snippets

Quantifying cab drivers' behavioral patterns

The increasing availability of GPS-based taxi trajectory data has led to an enormous amount of studies on human behavioral patterns, including both passengers and cab drivers (Liu, Kang, Gao, Xiao, & Tian, 2012; Tang, Liu, Wang, & Wang, 2015; Zhang, Xu, Tu, & Ratti, 2018; Zhao, Qin, Ye, Wang, & Chen, 2017). In this section, we mainly focus on studies on cab drivers' behavioral patterns based on GPS taxi trajectory data. The related studies are discussed as follows. For instance, Liu et al.

Study area and datasets

Wuhan — the capital city of Hubei Province — is a densely populated metropolitan area in central China. It has an area of 8594 km2 and a population of about 11 million as of 2017. The city is an economic, education, and cultural hub in central China. In this study, we focus on the area within the Wuhan Outer Ring Road, where the majority of taxi trips occurred (Fig. 1a). Traffic analysis zone (TAZ) is selected as the analysis unit in this work. TAZs are normally constructed based on the

The regularities of cab drivers' stay activities

In this subsection, we explore the regularities of cab drivers' stay frequency based on the rank-size distribution and probability distribution. Fig. 3 illustrates the rank-size distributions and probability distributions of cab drivers' stay frequency on weekday and weekend at the TAZ level. Stay frequencies on weekdays and weekends are obtained by calculating the average values of frequency on weekdays and weekends respectively. As shown in Fig. 3(a), stay frequency decays dramatically for

Discussion and conclusions

Understanding cab drivers' stay behaviors could benefit the planning and management of public facilities that are tied to taxis' daily operations. However, very limited efforts have been devoted to studying this particular aspect of driver behavior. To fill the research gap, this study investigates cab drivers' stay behaviors using a GPS-based taxi trajectory dataset collected in Wuhan, China. The stay points of cab drivers are extracted from the trajectory data. These activities are then

CRediT authorship contribution statement

Pengxiang Zhao:Conceptualization, Methodology, Formal analysis, Investigation, Writing - original draft, Writing - review & editing.Yang Xu:Conceptualization, Methodology, Investigation, Writing - original draft, Writing - review & editing.Xintao Liu:Conceptualization, Investigation, Writing - review & editing.Mei-Po Kwan:Conceptualization, Investigation, Writing - review & editing.

Acknowledgement

The authors would like to thank the anonymous reviewers and the editor for their valuable comments and suggestions on earlier versions of the manuscript. This research is jointly supported by the Hong Kong Polytechnic University Start-Up Grant (no. 1-ZE6P, no. 1-BE0J), the National Natural Science Foundation of China (no. 41529101, no. 41801372), and a grant from the Research Committee on Research Sustainability of Major RGC Funding Scheme of the Chinese University of Hong Kong.

References (62)

  • S. Newnam et al.

    Exploring differences in driving behaviour across age and years of education of taxi drivers in Addis Ababa, Ethiopia

    Safety Science

    (2014)
  • J. Tang et al.

    Inferring driving trajectories based on probabilistic model from large scale taxi GPS data

    Physica A: Statistical Mechanics and its Applications

    (2018)
  • J. Tang et al.

    Uncovering urban human mobility from large scale taxi GPS data

    Physica A: Statistical Mechanics and its Applications

    (2015)
  • J. Wang et al.

    Environmental influences on leisure-time physical inactivity in the us: An exploration of spatial non-stationarity

    ISPRS International Journal of Geo-Information

    (2018)
  • M. Wang et al.

    Spatial disparities of uber accessibility: An exploratory analysis in Atlanta, USA

    Computers, Environment and Urban Systems

    (2018)
  • W. Wang et al.

    A comparative analysis of intra-city human mobility by taxi

    Physica A: Statistical Mechanics and its Applications

    (2015)
  • Y. Wang et al.

    Detecting anomalous trajectories and behavior patterns using hierarchical clustering from taxi GPS data

    ISPRS International Journal of Geo-Information

    (2018)
  • Y. Xu et al.

    Human mobility and socioeconomic status: Analysis of Singapore and Boston

    Computers, Environment and Urban Systems

    (2018)
  • Y. Xu et al.

    Unravel the landscape and pulses of cycling activities from a dockless bike-sharing system

    Computers, Environment and Urban Systems

    (2019)
  • L. Yang et al.

    Scalable space-time trajectory cube for path-finding: A study using big taxi trajectory data

    Transportation Research Part B: Methodological

    (2017)
  • X. Yang et al.

    Spatial heterogeneity in spatial interaction of human movements—Insights from large-scale mobile positioning data

    Journal of Transport Geography

    (2019)
  • Z. Yang et al.

    Analysis of Washington, dc taxi demand using GPS and land-use data

    Journal of Transport Geography

    (2018)
  • D. Zhang et al.

    Understanding taxi service strategies from taxi GPS traces

    IEEE Transactions on Intelligent Transportation Systems

    (2015)
  • W. Zhang et al.

    The fitness-to-drive of shift-work taxi drivers with obstructive sleep apnea: An investigation of self-reported driver behavior and skill

    Transportation Research Part F: Traffic Psychology and Behaviour

    (2018)
  • X. Zhang et al.

    Do different datasets tell the same story about urban mobility—A comparative study of public transit and taxi usage

    Journal of Transport Geography

    (2018)
  • J. Zhao et al.

    What influences metro station ridership in China? Insights from Nanjing

    Cities

    (2013)
  • P. Zhao et al.

    Uncovering the spatiotemporal patterns of co2 emissions by taxis based on individuals’ daily travel

    Journal of Transport Geography

    (2017)
  • Z. Zheng et al.

    Scaling laws of spatial visitation frequency: Applications for trip frequency prediction

    Computers, Environment and Urban Systems

    (2017)
  • C. Brunsdon et al.

    Geographically weighted regression: A method for exploring spatial nonstationarity

    Geographical Analysis

    (1996)
  • R. Ewing et al.

    Travel and the built environment: A meta-analysis

    Journal of the American Planning Association

    (2010)
  • Y. Gao et al.

    Optimize taxi driving strategies based on reinforcement learning

    International Journal of Geographical Information Science

    (2018)
  • Cited by (29)

    • Understanding the daily operations of electric taxis from macro-patterns to micro-behaviors

      2024, Transportation Research Part D: Transport and Environment
    • Rural residential energy-saving in China: Role of village morphology and villagers’ daily activities

      2022, Journal of Cleaner Production
      Citation Excerpt :

      Huang et al. characterized user's energy-use behavior, considering the impact of space design on users' daily activities, and guided users to reduce high-energy-consuming spaces through building design (O'Brien et al., 2020). Analyzing the time-space behavior characteristics of community residents plays an important role in the planning and design of traffic networks (Steinberg, 2021; Yuan et al., 2018) and user behavior on road (Jabri et al., 2020; Zhao et al., 2020). Extending this concept, we investigate how and to what extent that village morphology and residents' demographic structure influence farmhouses’ occupancy rate and energy consumption for sustainable building designs.

    • Spatial variation of ridesplitting adoption rate in Chicago

      2022, Transportation Research Part A: Policy and Practice
      Citation Excerpt :

      In addition, the spatial distributions of standard residuals for four GWR models are also examined, as presented in Fig. 7. We can find that only a few of the local regression models fail the residual tests (the census tracts with over 2.5 times the standard deviation) (Zhao et al., 2020). From the results of OLS models (see Table 2), it can be seen that for built environment variables, subway station density is negatively correlated with the RAR in most periods except for weekday evening peak, frequency of transit has a negative association with the RAR.

    View all citing articles on Scopus
    View full text