Tourist Perception and Sentiment Analysis of Destination Image from Online Travelogues: A Guizhou Case Study

This study investigates 830 travelogues concerning Guizhou tourism sourced from the Ctrip platform. Employing text mining, semantic network analysis, LDA topic modeling, and SnowNLP sentiment analysis, it establishes a four-dimensional research framework to deconstruct Guizhou's tourism image across cognitive, emotional, and holistic dimensions. Based on the findings, targeted recommendations are proposed. In particular, the research highlights how user-generated content offers rich and authentic insights into tourist perceptions, which are often more dynamic and diverse than those derived from traditional surveys. The combination of semantic network analysis and topic modeling not only captures the structural characteristics of discourse but also reveals the thematic focus areas that shape Guizhou's destination image. Moreover, sentiment analysis provides a nuanced understanding of visitors' emotional experiences, ranging from satisfaction with natural scenery and ethnic culture to concerns over infrastructure and service quality. This multidimensional exploration contributes to a more comprehensive picture of Guizhou tourism and deepens the understanding of tourists' cognitive-affective evaluations.