부산대학교 도서관

The prompt acquisition and distribution of precise disaster-related information are crucial for minimizing damage during disasters. Then, the prompt acquisition and distribution of precise disaster-related information are crucial. Therefore, the usefulness of social media such as Twitter, for disaster mitigation is gaining worldwide notice. In this study, by extracting words representing areas (prefecture names and municipality names) from Japanese tweets posted during large-scale disasters occurred in 2018 in Japan (Osaka Northern Earthquake that occurred in June and Heavy Rain Disaster in western Japan that occurred in July), we analyze the attention areas (areas mentioned a lot in tweets). Moreover, we implement an application to visualize the tweets analysis result in the form of a choropleth map and co-occurrence network. By using the proposed visualization application, we can easily understand how the attention area changes over time.