Exploring nonlinear and interaction effects of urban campus built environments on exercise walking using crowdsourced...

Area of Science:

• Environmental Science
• Urban Planning
• Public Health

Background:

• University campuses are vital for promoting physical activity, yet the specific environmental factors influencing walking remain unclear.
• Understanding these mechanisms is crucial for designing healthier and more sustainable campus environments.

Purpose of the Study:

• To analyze the nonlinear and interactive effects of campus built environments on exercise walking using machine learning.
• To identify key environmental factors that promote walking activity among students, faculty, and communities.

Main Methods:

• Utilized crowdsourced exercise walking data and diverse campus characteristics.
• Applied XGBoost algorithm and SHAP (SHapley Additive exPlanations) for explainable machine learning analysis.
• Evaluated variable importance, nonlinear relationships, and interaction effects.

Main Results:

• Proportion of sports land, proximity to water bodies, and Normalized Difference Vegetation Index (NDVI) significantly impact walking.
• Identified distinct nonlinear effects, including fluctuated or U-shaped patterns, and strong synergistic interactions among environmental factors.
• Six campus area types showed notable influence on walking activity.

Conclusions:

• Campus built environments, particularly sports facilities, green spaces, and water bodies, have significant, interactive effects on walking.
• Optimizing the spatial configuration of these elements can maximize their positive impact on physical activity.
• Findings provide targeted recommendations for creating inclusive, health-promoting, and sustainable campuses.