No open-access location is known for this record yet.
No open-access location is known for this record yet.
Objective Outdoor jogging has been gaining popularity worldwide thanks to its various health, social and environmental benefits. Optimizing the design of urban built environment is an effective way to promote outdoor jogging activities for urban residents. To achieve this, it is necessary to clarify the association and interaction between urban built environment and outdoor jogging activities. However, the existing research mainly focuses on the independent effects of single factors on jogging activities, while neglecting the interactive effects between built environment factors and the spatio-temporal differences in their impacts. To this end, this research aims to analyze the association and interaction between urban built environment and jogging behavior, so as to provide a scientific basis for creating a human-friendly living environment. Methods Firstly, this research utilizes crowdsourced geographic information to measure outdoor jogging activities and urban built environments at a large scale. Specifically, jogging flow is calculated with jogging GPS trajectory data collected from the Edooon sport app. Eight built environment factors selected from the three dimensions of attractiveness, vibrancy and accessibility are calculated using multi-source spatial data, such as Baidu street view images, point of interest (POI) data and road data. Secondly, the Geographic Detector model is employed to investigate the associations and interactions between built environment factors and jogging behaviors. Thirdly, jogging flow is categorized into three types (morning peak, night peak and off-peak) at hourly scale, two types (weekdays and weekends) at daily scale, and four types (residential, commercial, industrial and recreational) based on land use type. Then, the variations in the aforesaid associations and interactions are explored and analyzed from the perspective of different time and different land use types with the Geographic Detector model. Results Empirical analysis is conducted with real jogging-related GPS trajectory data recorded by the sports app of 9,860 users in Chengdu City, China. Results show that population density, land use mix, and density of public transit station are core factors determining the spatial distribution of outdoor jogging. Across different time scales, the eight built environment factors, including green view index (GVI), normalized difference vegetation index (NDVI), park density, population density, land use mix, retail store density, road intersection density and bus stop density, have a significant impact on jogging activities. The explanatory power of different built environment factors varies significantly across different land use types. For example, the impact of park density on jogging is significantly greater in recreational land than in industrial or commercial land. Nevertheless, for industrial land, only three indicators, namely population density, land use mix, and accessibility, have significant impacts on outdoor jogging. Significantly, the explanatory power of interacted built environment factors for jogging flow is greater than that of single factors. This indicates that the differences in spatial distribution of jogging flow result from the combined effects of multiple built environment factors. The interaction between population density and land use mix explains the spatial variation of jogging activities to an extent of 50%. The interactive effects of some environmental factors on jogging may change over time and land use types. For instance, during morning peak period, outdoor jogging is more affected by the interaction of park density, GVI, and land use diversity in Chengdu. In contrast, outdoor jogging during the evening peak period is mainly affected by the interaction of population density, bus stop density, and land use diversity. The core interacting factors affecting the distribution of jogging flow in residential and commercial areas are population density and land use mix. In contrast, the core interacting factors in recreational and industrial areas are land use mix and park density. The interactive effects between population density/land use mix and other built environment factors may be synergistically enhanced, and the explanatory power of these core interaction factors varies over space and time. For example, after the interaction of two built environment factors, the explanatory power may increase by more than 10% in commercial land, 20% in industrial land, while only around 7% in residential land. These differential impacts will help planners formulate targeted design strategies for environmental intervention. Conclusion This research utilizes crowdsourced geographic information and a geographic detector model to establish multiple detection models, identifying the dominant built environment factors affecting outdoor jogging activities and quantitatively measuring the interactive effects of different built environment factors. Crowdsourced geographic information provides a new, extensive, and cost-effective means for measuring human activities and built environment characteristics. The interactive effects of multiple built environment factors on outdoor jogging activities are greater than those of single factors, exhibiting both dual-factor enhancement and nonlinear enhancement effects. Moreover, the interactive effects of built environment factors on outdoor jogging vary with time and land use types. This variability reflects the behavioral decisions of urban residents under the constraints of multiple scenarios, such as time (e.g., leisure time, commuting time), space (e.g., location), infrastructure (e.g., transportation infrastructure, sports facilities), and environment (e.g., visual environment, safety). A systematic analysis of the interactions between built environment and outdoor jogging may support health-oriented urban design.
Loading…
Stored only when both papers are in the corpus — 4 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.