Insights from negative reviews in high-ratings, and the associations between park attributes: A case study in Park City, Chengdu

Abstract

Urban parks, a public product and service, can provide ecological, social, cultural, and recreational resources essential for sustainable urban development. Park attributes are valuable indicators for evaluating parks, and park planning and management personnel must consider visitors’ perceptions of these attributes. However, previous studies often ignored or underestimated negative attributes, especially in high-rating reviews. Moreover, parks have multiple attributes, rather than a singular one, and there exist potential associations among these attributes. Here, we selected 63 urban parks in Park City, Chengdu, China, as a case study; we used online reviews from Ctrip and Dianping, and Latent Dirichlet Allocation (LDA) combined with a manual coding method to explore visitors’ perceptions (positive or negative) of each attribute in each review. Initially, we analyzed how park attributes affected overall dissatisfaction using multiple linear regressions, and counted the positive and negative attributes in each review rating. We then measured the association rules between various attributes and perceptions. We found that both positive and negative attributes (84.12% and 54.86%) mainly appear in high-rating reviews, while low ratings accounted for only 15.69% of negative attributes. Attributes in high-rating reviews were more generalized and had a more significant effect than in low-rating. Additionally, three association rules were identified, including cases where two positive attributes (11 instances), two negative attributes (6 instances), or one negative and one positive attribute appeared together (2 instances), all within the cultural and historical attributes category. This study’s outcomes contribute novel insights to the realm of sustainable urban park planning and management.