The Spatial Mechanism and Predication of Rural Tourism Development in China: A Random Forest Regression Analysis

Xishihui Du, Zhaoguo Wang, Yan WangISPRS Int. J. Geo Inf 2023 · DOI

TL;DR

This study investigates the spatial patterns and driving factors of rural tourism development in China, using a combination of spatial statistics and machine learning. By identifying key rural tourism villages and analyzing their relationship with various natural, social, and rural factors, the research provides a predictive model for identifying suitable areas for future rural tourism development, crucial for sustainable rural revitalization.

What Problem Does This Paper Address?

China's rural revitalization strategy heavily relies on developing rural tourism. However, the rapid growth of this sector has led to negative consequences, necessitating a more optimized and sustainable developmental framework. While previous research has explored the spatial distribution of rural tourism, there's a need for a deeper understanding of the underlying spatial mechanisms and a predictive model to guide future development. This paper aims to fill this gap by analyzing the spatial distribution of "key rural tourism villages" (designated by the government) and identifying the key factors that influence their location and potential for development.

Methods

The study employs a multi-faceted approach:

  • Spatial Cluster Analysis (HDBSCAN): This density-based clustering algorithm was used to identify distinct spatial clusters of key rural tourism villages, revealing patterns of concentration and dispersion across China.
  • Kernel Density Analysis: This technique was used to visualize the spatial density of key villages and influential factors, creating heatmaps that highlight areas of high and low concentration.
  • Bivariate Analysis (Pearson's Correlation): This statistical method was used to examine the linear relationship between village density and individual influential factors, quantifying the strength and direction of these associations.
  • Random Forest (RF) Regression: This machine learning algorithm was the core of the predictive modeling. It was used to:
  • Assess the relative importance of various influencing factors on village density.
  • Develop a predictive model to identify suitable zones for rural tourism development across China.

Datasets

The study utilized the following datasets:

  • Key Rural Tourism Villages: Data for 1399 key rural tourism villages were obtained from the Ministry of Culture and Tourism of the People's Republic of China. Their coordinates were acquired using the Baidu Map API.
  • Influential Factors: A comprehensive set of 12 factors were collected and processed, categorized into:
  • Natural Dimension: Digital Elevation Model (DEM), Natural Environment (Rivers), Air Quality (PM2.5).
  • Social Dimension: Tourism Development (Domestic Visitors), Tourism Potential (Prefecture-level Administration), Economic Development (Per Capita GDP), Economic Vitality (Night-time Light), Economic Potential (Population Density).
  • Rural Dimension: Livelihood Diversity (Ecosystem Services), Culture Resources (Intangible Heritage), Tourism Resources (A-class Scenic Spots), Rural Accessibility (Road Network Density).

Data sources included government websites, research platforms, and atmospheric composition analysis groups. Data were standardized to a 30x30 km fishnet grid using ArcGIS Pro 3.0 for consistent analysis.

Key Insights and Ideas

  • Concentrated but Dispersed: Key rural tourism villages in China exhibit a concentrated spatial distribution, particularly in the eastern and central regions (south of the Hu Line), but also show dispersed patterns in other areas.
  • Regional Disparities: Significant regional disparities exist, with the Yangtze Delta, Beijing-Tianjin area, and Yellow River basin showing high concentrations, while western and border regions face constraints.
  • Cultural and Tourism Assets are King: The most significant drivers of rural tourism village density are cultural resources, tourism resources, tourism potential, and rural accessibility. This highlights the importance of heritage, natural attractions, and ease of access.
  • Air Quality as a Draw? Interestingly, higher levels of PM2.5 (air pollution) were positively correlated with village density, suggesting a potential demand for rural escapes from polluted urban environments.
  • Predictive Power of RF: The Random Forest model proved effective in identifying key determinants and predicting suitable areas for rural tourism development, offering a valuable tool for policymakers.

Key Results

  • Cluster Analysis: HDBSCAN identified 26 distinct clusters, with the top three (Clusters 4, 2, and 1) accounting for over 32% of key villages, located in the Yangtze Delta, Yellow River region, and around Beijing, respectively.
  • Factor Importance: Random Forest analysis revealed that cultural resources (1.67), tourism resources (0.9), tourism potential (0.47), and rural accessibility (0.34) collectively explained 78% of the variance in village density.
  • Suitability Mapping: The study produced a map classifying areas into five tiers of suitability for rural tourism development, highlighting the Beijing and Shanghai urban circles as most favorable, with specific suitable zones identified in northern and western regions, while extensive areas in the northwest are deemed unsuitable.

Limitations and Future Work

  • Data Gaps: The study acknowledges that missing data in certain regions might impact the research outcomes.
  • Supply-Side Bias: The selection of influencing factors was primarily focused on supply-side aspects, with limited consideration for demand-side factors.
  • Future Research: The authors suggest exploring other machine learning algorithms beyond Random Forest for future investigations. They also recommend incorporating demand-side factors to provide a more holistic understanding of rural tourism development.

Summary updated

BibTeX
@article{du2023spatial,
  author = {Xishihui Du and Zhaoguo Wang and Yan Wang},
  title = {{The Spatial Mechanism and Predication of Rural Tourism Development in China: A Random Forest Regression Analysis}},
  journal = {ISPRS Int. J. Geo Inf.},
  year = {2023},
  doi = {10.3390/ijgi12080321},
}
Figure