Decoding the spatial experience of the Nanjing museum cluster: A computational aspect-based sentiment analysis (ABSA)

Abstract

Museums, as key public cultural spaces, involve a complex interplay between spatial organization, exhibition content, and visitor experience. Existing evaluation methods face a trade-off: post-occupancy evaluation offers depth but is costly and difficult to scale, while online review studies provide large samples but lack the granularity to assess specific spatial and experiential elements. To address this gap, this paper introduces Aspect-Based Sentiment Analysis (ABSA) to develop a museum spatial experience evaluation framework. Using 39,969 online reviews from six museums in Nanjing, the study constructs a dual-level dimension system covering Spatial Environment and Museum Experience. A key innovation is the development of a dual-source aspect lexicon, blending professional architectural and exhibition terminology with high-frequency expressions from visitor reviews. Through aspect identification and sentiment scoring, the method generates cross-museum indicators of aspect salience and sentiment orientation. Results show that visitor evaluations focus heavily on Place Attachment and Identity, as well as Interpretive Systems and Learning Outcomes. This pattern indicates that assessment is shaped by cultural identification, meaning-making, and physical conditions together. This approach converts fragmented online reviews into actionable evaluation evidence for spatial and operational improvements, supporting evidence-based interventions in wayfinding, ancillary space provision, interpretive systems, and overall visitor experience.