Trends of Social Interest in Disaster Evacuation and Sheltering with COVID-19: Through the Analysis of Web Articles Using Text Mining Method

Abstract

The authors employed a text mining method based on published web articles to analyze how social interest in disaster evacuation and sheltering with COVID-19 changed over time. To understand social interest, the authors divided the number of COVID-19 positive cases in Japan into five phases. The results revealed a vague concern about the need for measures taken by local governments in Phase I. Furthermore, there were descriptions of actual countermeasures and training based on the heavy rain in July 2020, typhoon No. 9 (Maysak), and typhoon No. 10 (Haishen) in Phase III. Finally, in Phase V, it was possible to grasp how the social interest shifted to specific content.