Welcome to Chinese Journal of Applied Ecology! Today is

Chinese Journal of Applied Ecology ›› 2026, Vol. 37 ›› Issue (7): 2329-2342.doi: 10.13287/j.1001-9332.202607.028

Previous Articles     Next Articles

Threshold effects of built environment and social media on positive sentiment of visitors in Beijing parks

KONG Lingqiang1, WEI Hui2, ZHOU Wenhao3, FANG Xuening3*, QIAO Jianmin4, ZHANG Hang1, FENG Lijuan4, ZHAO Wenwu5   

  1. 1School of Public Administration, Zhejiang University of Finance and Economics, Hangzhou 310018, China;
    2School of Public Administration, Hebei University of Economics and Business, Shijiazhuang 050061, China;
    3School of Environmental and Geographical Sciences, Shanghai Normal University, Shanghai 200234, China;
    4College of Geography and Environment, Shandong Normal University, Ji'nan 250358, China;
    5Faculty of Geographical Science, Beijing Normal University, Beijing 100875, China
  • Received:2026-03-03 Accepted:2026-06-01 Online:2026-07-18 Published:2027-01-18

Abstract: In the digital era, urban parks contribute to public well-being not only through ecological and spatial functions but also through the influence of social media within an “online-offline” coupled experience. Previous studies have largely emphasized the role of the built environment, but overlooked the threshold effects of both the built environment and social media on positive sentiment of visitors. This gap constrains the advancement of refined and digitally enabled park governance. Using 98 parks with high social media activity within Beijing's Fifth Ring Road as a case, we examined the nonlinear relationships and threshold effects linking the built environment, social media, and visitors' positive sentiment, by integrating multi-source data-including social media and remote sensing and employing machine learning approaches, such as random forest, mean decrease impurity (MDI), mean decrease accuracy (MDA), and Shapley additive explanations (SHAP). The results showed that the association between social media factors and visitors' positive sentiment was stronger than that between most traditional park built-environment factors and visitors' positive sentiment. The cumulative relative importance of social media factors across MDI, MDA, and SHAP reached 83.3%, 92.8%, and 82.4%, respectively. Both built environment and social media factors exhibited multiple nonlinear thresholds. Specifically, the median threshold of Dazhong Dianping rating was 4.88, above which positive sentiment increased sharply. The median threshold for park area was 47.12 hm2, beyond which the positive correlation tended to level off. The median threshold for average housing price was 135400 yuan·m-2, above which its relationship with positive sentiment shifted from neutral to negative. Our results highlighted the critical role of social media in shaping park visitors' emotional experiences and revealed the complex nonlinear interactions between digital and physical environments. For urban parks with high social media engagement, park governance could shift towards precision-oriented approaches that prioritize the management of digital word-of-mouth and the optimization of resource allocation, thereby enhancing the high-quality, efficient, and equitable provision of green space services and supporting people-centered sustainable urban development.

Key words: visitors' positive sentiment, urban park, social media, threshold effect, digital governance, sustainable development