Discovering spatiotemporal patterns of human outdoor activities with crowdsourced trajectory data

Abstract

The growing emphasis on human physical and mental well-being has led to a surge in outdoor activities, underscoring the need to explore outdoor activity patterns. As a specific type of human activities, outdoor activities may exhibit distinct patterns, especially statistical regularities, which diverge from human routine activities such as commuting, shopping, and working. While existing studies primarily focus on the spatiotemporal patterns of outdoor activities, a more comprehensive investigation that integrates the statistical laws, which have been widely explored for human daily activities, remains limited. To bridge this gap, this study systematically explores the patterns of three representative outdoor activities (i.e. climbing, hiking, and running). Specifically, we examine the underlying statistical mechanisms of activity distances and durations that shape outdoor behaviors, investigate temporal rhythms and user preferences across daily, weekly, and monthly scales, and identify statistically significant hotspots of arbitrary shapes, characterizing their spatial distributions and seasonal variations. A case study using crowdsourced outdoor activity data from four cities in China revealed as follows: (1) Outdoor activities exhibited exponential or lognormal distributions in distances and durations, which differs from scale-free distributions commonly observed in human daily activities. The distances and durations of climbing and hiking activities are constrained by the accessible climbing and hiking routes in different cities, while running activities in four cities consistently follow lognormal distributions with multi-peaks observed in distance distributions. (2) Temporal patterns varied across outdoor activities and time scales, reflecting distinct temporal dynamics and user preferences. (3) The spatial distributions of outdoor activity hotspot areas and routes differed significantly across climbing, hiking, and running activities and exhibited distinct seasonal variations. These findings not only complement human mobility theory by expanding its scope to outdoor contexts but also contribute to the understanding of outdoor activity patterns and have practical implications for urban planning and policymakers.

Keywords

• Outdoor activity
• spatiotemporal patterns
• statistical law
• hotspot
• crowdsourced trajectory data
• behavior preference