Deciphering socio-spatial drivers of visitor sentiment in cultural heritage tourism: a text mining and XGBoost-based empirical study of Go Seigen in Fuzhou, China

The tension between rapid urbanization and cultural tourism heritage development presents persistent global challenges, particularly for celebrities, which often remains isolated from the urban socio-spatial fabric. This study explores how tourism heritage interacts with spatial quality and visitor sentiment, using geotagged socio-spatial multi-source data in Go Seigen’s cultural legacy in Fuzhou, China. Findings via regression and eXtreme Gradient Boosting (XGBoost) models indicate that tourism heritage sites exhibit limited socio-spatial implications in the digital era, reflected in declining social media check-ins and dispersed emotional expression. Spatial context, including cultural and leisure, and public facility POIs, has a stronger influence on visitor sentiment than heritage itself. Social media interaction density and POI distribution are dominant drivers, collectively boosting sentiment by 25%. This study advances theoretical discourse by integrating digital engagement with spatial enhancement, revealing that spontaneous digital interaction alone is insufficient to sustain tourism heritage management.