Quantifying the Relationship Between Public Facility Distribution and Regional Sentiment in Japan Using Twitter Data

Public facilities significantly shape urban livability, but planners rarely quantify their psychological impact. This study investigates how the spatial distribution of public facilities relates to regional sentiment, using geotagged Twitter data and facility information across Japan. We analyzed over 51 million tweets from 2022, applying lexicon-based sentiment analysis (ML-Ask) and aggregating scores at the 1 km 2 grid level. Regression analysis revealed that libraries and religious sites (temples/shrines) were positively associated with sentiment, while medical and welfare facilities showed negative correlations. We further applied k-means clustering to identify patterns in facility composition and corresponding sentiment. Clusters enriched with cultural or educational amenities tended to have higher sentiment scores, while those dominated by healthcare and welfare facilities had comparatively lower sentiment scores. These findings underscore the importance of facility composition in shaping the perceived emotional atmosphere of regions. Rather than focusing solely on increasing individual facility counts, considering the balance of cultural, educational, and social service facilities may offer more holistic benefits in spatial planning and community development.