Using social media data to improve specific urban park activities: the case of park camping

Emerging urban nature-based leisure practices in public parks require activity-oriented evidence that can identify experience gaps, prioritize improvement needs, and clarify feasibility boundaries under governance constraints; this study develops an interpretable workflow using urban park camping as a representative case. We collected 5449 Little Red Book posts, retained 4,840 valid posts after cleaning, and extracted 71 camping-related topics via LDA with TF-IDF weighting; keyword-level sentiment evaluation and ODI opportunity scoring were then applied to prioritize underserved topics, and semi-structured expert interviews provided cross-actor validation and feasibility clarification. Across all topic keywords, the mean sentiment score was 0.654, indicating generally positive evaluations, while negative evaluations concentrated on a limited set of high-salience elements within topics, enabling targeted diagnosis beyond theme-level summaries. Six high-opportunity topics were identified as priority areas for improvement, and integrating within-topic weight-sentiment configurations, opportunity scores, and expert evidence revealed three linked mechanisms explaining persistent gaps in park camping: misalignment between activity demand and feasible supply, scale differences between comfort-oriented environmental evaluation and cumulative ecological constraints, and conflict amplification through shortcomings in basic facilities and reservation management. Methodologically, the study operationalizes transparent prioritization by combining importance- and satisfaction-related signals and by locating dissatisfaction drivers at the keyword level; practically, it provides a structured basis for targeting peak-sensitive bottlenecks and communicating feasibility boundaries for boundary-sensitive demands in urban parks.