打印

作者

Xiaoquan Wang;Chunfu Shao;Chaoying Yin;Ling Guan

来源

CITIES,Vol.110,Issue1,Article 103048

语言

英文

关键字

Built environment;Relative importance;Interactive effects;Obesity;Different geographic scales;Machine learning method

作者单位

Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport, Beijing Jiaotong University, Beijing, China;College of Automobile and Traffic Engineering, Nanjing Forestry University, Nanjing, China;Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport, Beijing Jiaotong University, Beijing, China;College of Automobile and Traffic Engineering, Nanjing Forestry University, Nanjing, China

摘要

One common goal of urban planning is to promote health, where the built environment (BE) is believed to play an essential role. This study provides new evidence for the links between body mass index (BMI) and the BE at different geographic scales in a unified framework. Based on a sample of 10,962 individuals across 95 cities in China, this study explores the collective effects of the BE attributes at neighborhood and city levels on BMI using a gradient boosting decision tree (GBDT) approach. Moreover, whether the relationships between the city-level BE attributes and BMI vary across social groups is revealed. The results indicate that the city-level BE attributes have larger collective effects on BMI than the neighborhood-level BE attributes. Specifically, population size is the greatest contributor to predicting BMI, with a relative importance of 13.92% among all influential factors. Moreover, it is found that the effects of city-level BE attributes vary across different social groups. These findings suggest it is of vital importance for urban planners to coordinate both neighborhood- and city-level BE planning to deter the risks of obesity, and urban planning can be more effective if it is implemented considering local socio-demographic characteristics.