ユーザー生成コンテンツを活用した日本人と外国人の観光行動分析: 世界文化遺産の村、白川郷を事例に

This study examined the usefulness of text mining in analysing tourism behaviour by comparing and analysing electronic word of mouth (eWOM) concerning Shirakawa-go from both Japanese and foreign individuals posted on a travel review website. Comparison of Japanese and English eWOM, as clustered by the latent Dirichlet allocation, revealed a certain degree of commonality in the content of both versions on several topics related to sightseeing behaviours in the village ─ such as eating, walking and touring, and taking photographs ─ while significant differences were found between the two versions on the topic of World Heritage sites. Negative reactions to the rapid increase in foreign tourists appeared in the Japanese version, whereas the characteristic words in the English version showed that the target site was consumed as a place to enjoy a unique folk culture. Differences also emerged between the two versions on the topic of means of transport, reflecting the number of means of transport to Shirakawa-go available to both parties. The cooccurrence network analysis, stratified by the subjects’ place of departure, also provided findings that corroborate the existing interview survey results, and showed its effectiveness as a means of reinforcing the findings from an existing qualitative survey.