A method to characterize the social cascading damage processes of disasters using media information

Constantly advancing media information is a key data source to characterize the social cascading damage processes following natural hazards. However, media information tends to include a large sample size but low information density. In consideration of these properties, the aim of this study is to develop a new method for media-based information characterizing social cascading damage processes. In developing the method, a network theory framework was constructed to systematically integrate media information and its characterization. The method has two analytical components: a disaster damage network systematically inputting media information and network analysis using the concept of degree centrality. The developed method was applied to the record-breaking 2018 heavy rain disaster in western Japan, employing newspaper articles as media information sources. The study identified the critical disaster events and their relationships. This case study demonstrates that our method will benefit policymakers by providing them with potential fundamental information to support disaster management.