Assessing Spatial Heterogeneity in Urban Park Vitality for a Sustainable Built Environment: A Case Study of Changsha

Liwei Qin, Wenke Zong, Kai Peng, Rongpeng ZhangLand 2024 · DOI

TL;DR

This study tackles the challenge of understanding urban park vitality by moving beyond "average" effects to reveal how different factors influence park usage across diverse urban areas in Changsha. Using a novel Multi-scale Geographically Weighted Regression (MGWR) model and big data, it uncovers significant spatial variations in park vitality, highlighting the importance of local context for effective urban planning.

What Problem Does This Paper Address?

Urban parks are crucial for sustainable city development, offering environmental benefits and spaces for recreation and social interaction. However, simply measuring the quantity of green space isn't enough; understanding how these spaces are used and why their usage varies is key to improving public space quality and urban livability. Traditional methods for assessing park vitality often rely on surveys or global models that can mask crucial spatial differences. This paper addresses this gap by investigating the spatial heterogeneity of park vitality, meaning how the influence of various factors on park usage changes across different locations within a city. The authors aim to provide a more nuanced understanding of what drives park vitality, moving beyond generalized insights to inform localized planning strategies.

Methods

The study employs a three-step analytical framework:

  1. Sample Generation and Variable Determination: This involves collecting spatial location data for urban green spaces (UGS) and identifying key influencing factors (independent variables) and the dependent variable (park vitality).
  2. Key Indicator Calculation (Global Analysis): An Ordinary Least Squares (OLS) regression is used to establish a baseline understanding of the relationships between variables and to identify and address potential multicollinearity issues among the independent variables. This step provides a global perspective before diving into spatial variations.
  3. Spatial Heterogeneity Analysis: This is the core methodological contribution. The study utilizes two advanced spatial regression models:
  4. Geographically Weighted Regression (GWR): This model allows for local regression, meaning it estimates relationships at different locations, accounting for spatial non-stationarity.
  5. Multi-scale Geographically Weighted Regression (MGWR): This is the novel approach. Unlike GWR, which uses a single bandwidth for all variables, MGWR allows each independent variable to have its own unique bandwidth. This is crucial because different factors (e.g., proximity to residential areas vs. availability of sports facilities) might influence park vitality at different spatial scales. MGWR is thus better equipped to capture the nuanced spatial heterogeneity of these influences.

Datasets

The study utilizes the following datasets:

  • Baidu Huiyan Vitality Data: This dataset provides real-time population distribution data derived from mobile phone location requests, serving as a proxy for park vitality (check-ins).
  • Amap Geospatial Information: This includes Area of Interest (AOI) and Point of Interest (POI) data, which are used to identify and characterize internal and external factors influencing park vitality (e.g., residential communities, commercial facilities, transportation hubs, sports facilities).

Key Insights and Ideas

  • Spatial Heterogeneity is Paramount: The most significant takeaway is that factors influencing park vitality are not uniform across the city. What makes a park vibrant in one area might have a different effect elsewhere.
  • Context Matters for Planning: Generic urban planning strategies for parks are insufficient. Understanding the unique socio-economic and spatial characteristics of different urban zones is crucial for effective park design and management.
  • MGWR Outperforms Traditional Models: The MGWR model, by allowing variable-specific bandwidths, provides a more accurate and nuanced understanding of spatial relationships compared to OLS and even GWR. This is critical for dissecting complex urban phenomena.
  • Internal vs. External Factors Interact Differently: The study highlights how internal park features (like business amenities) and external factors (like residential proximity) have varying impacts depending on their spatial context.

Key Results

  • Vitality Hotspots and Coldspots: The study identified distinct patterns of park vitality, with areas west of the Xiangjiang River near commercial and residential developments showing higher vitality, while the northwest region exhibited lower vitality, potentially due to later development stages and a higher concentration of industrial parks.
  • Internal Factors:
    • Business and Public Toilets: Show a positive correlation with park vitality, particularly in the southwestern region.
  • Open Space: Exhibits a negative correlation with park vitality, suggesting that an overabundance of open space might lead to individual dispersal and reduced vibrancy.
  • Sports Facilities and Entrance/Exit Points: Show varied spatial distributions of influence, with sports facilities being more impactful in the central area and entrance/exit points in the northwest.
  • External Factors:
    • Residential Communities, Companies, and Public Facilities: Generally show a positive correlation with park vitality. Residential communities, in particular, had the most substantial impact, with higher influence in the west.
  • Transportation Facilities: Seem to have an inhibitory effect on park vitality, especially in areas with significant job-housing segregation and long commutes.

Limitations and Future Work

  • Data Limitations: The study acknowledges potential inaccuracies in spatial vitality data due to errors in location features and possible underrepresentation of specific demographic groups (e.g., older adults, children).
  • Future Work: The authors recommend exploring and implementing novel data collection methods to address these limitations and ensure more comprehensive and representative datasets for future park vitality assessments. They also suggest tailoring strategies based on specific human resource needs in different areas.

Summary updated

Abstract

In the realm of sustainable city development, evaluating the spatial vitality of urban green spaces (UGS) has become increasingly pivotal for assessing public space quality. This study delves into the spatial heterogeneity of park vitality across diverse urban landscapes at a city scale, addressing limitations inherent in conventional approaches to understanding the dynamics of park vitality. Leveraging geotagged check-in data from 65 parks in the study case of Changsha City, a quantitative analysis was undertaken to assess spatial vitality. The investigation incorporated data concerning internal and external factors influencing park vitality, employing the Multi-scale Geographically Weighted Regression (MGWR) model to dissect nuanced spatial heterogeneity. The research uncovers notable spatial discrepancies in factors influencing park vitality across diverse urban areas, emphasizing the reliance on adjacent residential communities and internal commercial amenities provision. These dependencies correspond with economic development differences among urban locales, revealing distinct geographic trends. This study has a novel perspective and methodology for investigating urban park vitality, providing significant insights for urban green space planning and management. It emphasizes the necessity of acknowledging spatial diversity in urban park planning and design by incorporating the distinct socio-economic characteristics of each urban zone, which is crucial for both urban planners and policymakers.
BibTeX
@inproceedings{qin2024assessing,
  author = {Liwei Qin and Wenke Zong and Kai Peng and Rongpeng Zhang},
  title = {{Assessing Spatial Heterogeneity in Urban Park Vitality for a Sustainable Built Environment: A Case Study of Changsha}},
  booktitle = {Land},
  year = {2024},
  doi = {10.3390/land13040480},
}
Figure