Objective The emergence of volunteered geographic information (VGI) has provided new possibilities for accurately evaluating the use efficiency of physical activity in the built environment. To systematically analyze the association between urban neighborhood environments and mobile physical activity, a large number of existing empirical researches need to be systematically summarized and reviewed. However, there is a lack of systematic review and analysis based on VGI data, which makes it difficult to pinpoint information on the location and spatial distribution of individual physical activities, and to reflect the built environment preferences of mobile physical activities. Understanding the association between physical activity and the characteristics of built environments is of great significance for proactive intervention in public health.
Methods To systematically validate the use efficiency of mobile physical activity in the built environment, this research screens 31 academic papers with complete descriptive statistical reporting from Web of Science database and other databases using such keywords as volunteered geographic information, physical activity, and environmental characteristic. The literature included in the meta-analysis is systematically sorted out with the main information extracted, which can be divided into the following five aspects: 1) Article information: title, author, and year of publication; 2) basic sample information: research area and scale, and sample size; 3) analysis of the research method: type of regression model, basic unit of spatial analysis, extent of buffer zone, implementation of descriptive statistics or not, control of socioeconomic attributes or not, etc.; 4) information on dependent variables: dimensions of physical activity indicators (intensity, diversity and suitability), type of mobile physical activity, and source of physical activity data; 5) information on independent variables: selection of environmental indicators (physical, social and perceived indicators) and data sources thereof. Based on this, a quantitative meta-analysis is conducted on the results of the aforesaid academic papers.
Results The research results show that the natural environment, built environment, social environment and subjectively perceived environment all have consistent and significant associations with mobile physical activities, with the degree of association varying according to the type of physical activity. As for the natural environment, top-down greenness assessed by normalized difference vegetation index (NDVI) or green space density has the strongest positive correlation with physical activity. Eye-level greenness is also highly correlated with cycling activity. The existence of aquatic environment is highly correlated with cycling and running activities. Tree density, air quality and river band length have had positive associations with running activity. Greenway density is found positively associated with general physical activity. As for the traffic-related environment, bike lane density, curbstone presence, traffic-related accident density, and public transport node density are highly associated with cycling activity. Road density also has a consistent and significant positive association with mobile physical activity, whereas the provision of amenities and the width of sidewalks only positively contribute to walking activity. The connectivity of street network is positively associated with general physical activities. As for land use and other built environments, residential density and open space density have a significant and positive association with walking, running and general physical activity, but not with cycling. Land use mix is highly associated with both cycling and walking activities. Terrain slope is only highly associated with cycling activity, while campus density and facility density are only positively associated with running activity. Facility diversity has a moderate association with running activity. Restroom density is positively associated with walking, running and cycling activities. Compared to the objective built environment, there are fewer researches on the impact of social and subjectively perceived environments on physical activity. Only population density is highly associated with walking behavior. The measure of subjectively perceived environment mainly involves environmental safety perception, environmental vitality and environmental richness, of which only environmental safety perception is highly associated with running activity.
Conclusion The large-volume, multi-scale and high-precision volunteered geographic information on physical activity can help objectively grasp the distribution of physical activities in urban neighborhoods, and compare the access patterns and use efficiency of physical activity in different built environments at various spatio-temporal scales, thus making it possible to build a more scientific and rational human environment that can promote healthy behaviors. Findings based on the meta-analysis may provide an empirical model for use efficiency prediction for urban planners and policymakers optimizing and constructing new physical activity intervention facilities. Future research may use VGI from different periods to longitudinally explore the spatio-temporal differentiation of mobile physical activities and their associations with built environment factors, and analyze the impact of various types of built environment factors on physical activity at different stages of urban development to build a living environment that can scientifically promote healthy behaviors.
DownLoad: