Comparing Urban Resilience Strategies: Leveraging Large Language Models for Comparative Analysis

• In Event: Governance in Action: Comparative Studies of Bureaucracy and Policy

Abstract

As climate threats intensify, cities are playing an increasingly central role in resilience planning. The Resilient Cities Network (RCN) offers a valuable platform for sharing strategies, creating opportunities to explore both commonalities and differences in local approaches. This paper explores the use of large language models (LLMs) for the comparative analysis of urban resilience documents. Drawing on resilience strategies from a diverse sample of cities within the RCN, the study uses an LLM-assisted workflow to identify thematic similarities, policy emphases, and linguistic framings across cases. By combining qualitative content analysis with the pattern recognition and synthesis capabilities of LLMs, the research reveals both common priorities—such as infrastructure adaptation and social equity—and distinctive local adaptations shaped by governance, and resource constraints. The paper reflects on the methodological advantages and challenges of using LLMs for political science research, including questions of transparency, validity, and interpretability. Ultimately, the study demonstrates how LLMs can augment traditional methods in policy analysis and contribute to a more scalable and systematic understanding of global urban resilience.

Author

• Kristen Hudak Rosero, Wentworth Institute of Technology