Computational Text Analysis of a Scientific Resilience Management Corpus: Environmental Insights and Implications

Abstract

Resilience is a multifaceted concept describing the ability to cope with change or disruption. Its importance in the era of emergency preparedness and response, combined with its multidisciplinary attributes, have led researches to study similarities and differences in the meaning of resilience across various fields. A systematic literature review, conducted in the field of resilience management by the DARWIN project, yielded a scientific corpus of 419 articles. In the present study, automated text-analysis approaches were used to investigate this corpus and generate insights, aiming at understanding resilience management. Three complementary computational analyses were employed: (a) topic modeling to understand the different topics or fields discussed in the articles; (b) concept maps to provide a synthetic view of key concepts in the domain and their relations; (c) psycho-linguistic analysis to identify significant psychological categories addressed in the corpus. The topic model identified four key topics: Environmental/Socioecological aspects, Organizational/Operational aspects, Health, and Infrastructure/Resource Management. The concept map recognized concepts at a finer granularity level and depicted them into five main clusters with relations between them, reflecting key dimensions leading to resilience management. The psycho-linguistic analysis highlighted the importance of psychological processes within resilience management. This study identified important aspects that need to be addressed when designing resilience management frameworks, such as rehabilitation period and the role of public.

Keywords: concept maps, LIWC, NLP tools, resilience, resilience management, topic modeling