摘要：Abstract Mining hotel social sensing data and analyzing its spatial and temporal characteristics can provide decision support for hotel managers. Present research on this topic is limited to the overall hotel industry and text mining. Here, we first obtain POI and reviews for star-rated hotels in Nanchang from 2018 to 2021. Secondly, the hotel POI (Point of Interest) is combined with the sentiment value of customer reviews. Finally, comparative analysis and topic mining of Spatio-temporal aspects of text reviews of different star-rated hotels are conducted using sentiment analysis, spatial analysis, and thematic social network analysis. Results show that: (1) Hotel star rating and hotel review sentiment value are significantly positively correlated. The seasonal trends of different star rating hotel sentiment values are similar, but are highest in summer and lower in autumn; (2) The highest sentiment value is seen for friends’ outings and the lowest is for business trips; (3) Customer reviews of star-rated hotels focus on three aspects: facilities, service, and location. Three-star hotels focus on the stay experience, while four-star hotels focus on the breakfast situation. Exploring hotel social sensing data can intuitively illustrate hotel selection’s behavioral patterns and spatial-temporal characteristics. The methods of this study can expand the application of social sensing data in different fields, such as the tourism and restaurant industries.
