This study analyzed Sense of Place (SOP) from the perspective of place branding, in light of the aim in Japan to build a sustainable social system utilizing regional cooperation. To date, most SOP analyses have been influenced by subjectivity, but AI is increasingly being used for this analysis. This study focused on the Tsubame-Sanjo region, a representative example of an inter-regional brand. Data were collected for SNS postings about the region, which were then analyzed using machine learning to evaluate the feelings of residents and tourists toward the region. A transformer-based Luke model was used to analyze the emotions. Eight basic emotions, including joy, expectation, and surprise, were identified from the contents of the posts, and this allowed identification of changes in emotions associated with Tsubame-Sanjo. The analysis also showed that events were associated with expectation and joy, and that the effects of the COVID-19 pandemic were also reflected in the emotions. This study proposes a method for measuring the effects of events and strengthening place brand power through use of SNS data, and we suggest that this method is applicable to other regions.
View full abstract