PROGRESS IN GEOGRAPHY ›› 2023, Vol. 42 ›› Issue (1): 79-88.doi: 10.18306/dlkxjz.2023.01.007

• Articles • Previous Articles     Next Articles

Nonlinear associations between urban vitality and built environment factors and threshold effects: A case study of central Guangzhou City

WANG Chenggang1(), WANG Bo2,3,*(), WANG Qizhi2, LEI Yaqin2   

  1. 1. Guangzhou Urban Planning & Design Studio, Guangzhou 510030, China
    2. School of Geography and Planning, Sun Yat-sen University, Guangzhou 510275, China
    3. Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519000, Guangdong, China
  • Received:2022-06-03 Revised:2022-08-20 Online:2023-01-28 Published:2023-03-28
  • Contact: WANG Bo;
  • Supported by:
    National Natural Science Foundation of China(42271210);National Natural Science Foundation of China(41901191);Natural Science Foundation of Guangdong Province(2022A1515011572);Guangzhou Basic and Applied Basic Research Foundation(202102020795)


The relationship between urban vitality and the built environment has always been a subject of intense research in urban geography and planning. Traditionally, global regression techniques have been mostly developed to analyze their quantitative relationship based on a linear consumption. However, the results from existing studies are rather mixed, indicating that their relationship may not be global. Therefore, it is necessary to explore the local characteristics of the associations between urban vitality and the built environment. Based on a collection of multi-source datasets including Baidu Heat Map data, building data, road network data, and point of interest data in central Guangzhou City—including Liwan, Yuexiu, Tianhe and Haizhu districts, this study applied the gradient boosting decision tree model to unveil the nonlinear associations of built environment characteristics (including intensity, function, and accessibility) with urban vitality and threshold effects. The differences in the impacts of the built environment between daytime and nighttime on weekdays have also been examined. The results show that: 1) During both the daytime and the nighttime, floor area ratio contributes the greatest to urban vitality, followed by the intensity of recreation and public transport facilities. Compared to diversity, reasonable development intensity, concentration of leisure and office facilities, and public transport oriented development have larger collective contributions to urban vitality. 2) Although differences exist between daytime and nighttime, all built environment variables have nonlinear associations with urban vitality. The threshold value and gradient of key built environment variables are recognized in the nonlinear shapes of associations. Urban planners and local governments are recommended to meticulously disentangle the complicated built environment associations to make informed and targeted interventions for fostering and maintaining urban vitality.

Key words: urban vitality, built environment, nonlinear associations, threshold effect, machine learning, Baidu Heat Map