What Drives the Spatial Heterogeneity of Urban Leisure Activity Participation? A Multisource Big Data-Based Metrics in Nanjing, China

Shaojun Liu, Xiawei Chen, Fengji Zhang, Yiyan Liu, Junlian GeISPRS Int. J. Geo Inf 2023 · DOI

TL;DR

This paper tackles the challenge of understanding why people participate in urban leisure activities in different places. By integrating mobile phone data, geospatial information, and street view imagery, the researchers developed a novel framework to quantify leisure activity participation and identify its key drivers, revealing that factors like resource availability, accessibility, and even streetscape aesthetics play crucial roles, with variations across different leisure activity types.

What Problem Does This Paper Address?

As cities grow, understanding how residents use urban spaces for leisure is crucial for improving quality of life and urban planning. While previous studies have looked at urban leisure spaces, a significant gap exists in accurately measuring actual leisure activity participation at a city scale and understanding the diverse factors that drive it. Existing methods often rely on less precise data or fail to differentiate between various leisure activities, leading to a one-size-fits-all approach. This paper aims to bridge this gap by developing a robust method to quantify urban leisure activity participation (LAP) and, more importantly, to uncover the complex interplay of internal and external factors that shape its spatial distribution.

Methods

The researchers propose a novel, multi-source big data framework. First, they use mobile phone signaling (MPS) data to construct "stop chains" representing user movements and then employ machine learning (specifically, Random Forest, which achieved the highest accuracy) to classify these stops into activity types, including leisure. This allows for the estimation of LAP.

Next, to understand the drivers of LAP, they construct a comprehensive index system. This system quantifies both "internal" factors related to the leisure regions themselves (like density, diversity, and richness of leisure resources, as well as subjective perceptions of streetscape quality derived from street view images) and "external" factors (such as traffic accessibility, surrounding population density, and competition).

Finally, Ordinary Least Squares (OLS) regression is used to model the relationship between these driving indicators and the estimated LAP. To ensure model reliability, they also employ Variance Inflation Factor (VIF) to check for multicollinearity.

Datasets

The study utilizes several key datasets:

  • Mobile Phone Signaling (MPS) Data: Approximately 30 million subscribers' data from April 2019, providing location and timestamps for user activities.
  • Urban Geospatial Big Data: This includes Points of Interest (POI) for leisure venues, Areas of Interest (AOI), urban road networks, and building footprints, primarily sourced from Gaode Maps.
  • Web-Crawled Data: Dianping POIs were used to assess consumption levels.
  • Baidu Maps Street View Image (BMSV) Data: A substantial collection of street view images from multiple angles, used to evaluate subjective perceptions of street quality.

Key Insights and Ideas

  • Weekends are King: Leisure activity participation is significantly higher on weekends, with distinct daily peaks, highlighting the importance of understanding weekend urban dynamics.
  • Beyond Simple Density: The study moves beyond just counting leisure venues. It emphasizes that the diversity and richness of leisure resources, along with subjective perceptions of the environment (like greenness and walkability), are critical drivers of LAP.
  • Activity-Specific Drivers: A major contribution is the finding that different types of leisure activities (sports, sightseeing, culture, recreation) are driven by distinct factors. For instance, sports participation is influenced by street greenery and enclosure, while sightseeing relies more on transportation convenience and service facility abundance.
  • Streetscape Matters: Subjective perceptions of the street environment, captured through street view imagery, play a role, particularly for sports and cultural activities, suggesting that aesthetics and perceived safety contribute to engagement.
  • Population Density is Key: Higher surrounding population density generally correlates with increased leisure participation, underscoring the importance of accessible leisure options in densely populated areas.

Key Results

  • The Random Forest model achieved a high accuracy of 92% for classifying user activity types.
  • The OLS model for overall LAP explained 52.3% of the variance, with density, diversity, and richness of leisure resources, traffic accessibility, and surrounding population density being statistically significant drivers.
  • For specific activity types, the sports model showed the highest explanatory power (R-squared of 0.8571), followed by sightseeing (0.6735).
  • The study identified that while some factors like resource conditions are important for sports, sightseeing, and recreation, sports activities require a more balanced allocation of resources, whereas tourism and recreation benefit more from abundance.

Limitations and Future Work

  • Single City Focus: The study is based on Nanjing, China. The authors acknowledge that future research should incorporate other cities to compare LAP patterns and evaluate the adaptability of their proposed indicator systems.
  • Subjective Perception Enhancement: The current subjective perception assessment relies on street view images. Integrating assessments of indoor and outdoor environments within leisure areas could further enhance the understanding of leisure-driven mechanisms.
  • Data Privacy: The raw MPS data is not publicly available due to privacy concerns, which is a common limitation in this type of research.

This research provides a valuable framework and actionable insights for urban planners and policymakers to better understand and shape urban leisure experiences, ultimately contributing to more vibrant and livable cities.

Summary updated

BibTeX
@article{liu2023what,
  author = {Shaojun Liu and Xiawei Chen and Fengji Zhang and Yiyan Liu and Junlian Ge},
  title = {{What Drives the Spatial Heterogeneity of Urban Leisure Activity Participation? A Multisource Big Data-Based Metrics in Nanjing, China}},
  journal = {ISPRS Int. J. Geo Inf.},
  year = {2023},
  doi = {10.3390/ijgi12120499},
}
Figure