Disaster Records as Data: Natural Language Processing, Geocoding and Spatial Analysis of Hydrometeorological Occurrences in the São Paulo Metropolitan Region.

• Grant number: 26/05933-5
• Support Opportunities: Scholarships abroad - Research Internship - Post-doctor
• Start date: September 01, 2026
• End date: March 31, 2027
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Raquel Rolnik
• Grantee: Ulises Rodrigo Magdalena
• Supervisor: Keiller Nogueira
• Host Institution: Faculdade de Arquitetura e Urbanismo (FAU). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: University of Liverpool, England
• Associated to the scholarship: 25/02462-9 - Critical cartography: articulating multi-scalar and interdisciplinary cartographic methods, BP.PD

Abstract

This project proposes a Research Internship Abroad (BEPE/FAPESP), which will be undertaken as a visiting scholar at the University of Liverpool's School of Computer Science and Informatics under the supervision of Professor Dr Keiller Nogueira. During the six-month internship, the theoretical and analytical framework developed in my postdoctoral research project, "Critical Cartography: the articulation of multiscalar and interdisciplinary cartographic methods.", will be advanced. The central objective is to consolidate a systematic, replicable methodological framework for transforming unstructured disaster reports into structured, geocoded data using text mining and natural language processing techniques. The internship will refine procedures for extracting, standardising and spatially integrating information from the "Sistema Integrado de Defesa Civil do Estado de São Paulo (SIDEC)" database. This will ensure the identification and classification of hydrometeorological events and technological incidents recorded between 2014 and 2025 are conducted with methodological rigour. Particular emphasis will be placed on developing reproducible workflows for attribute extraction, typological standardisation, geocoding and validating classification outputs. Engaging with international research in computational modelling, geospatial analysis and disaster risk assessment will strengthen the integration of data science and critical urban studies. In collaboration with researchers at the University of Liverpool, the project will examine how disaster records contribute to the delineation of territories designated as "at risk" and to the production of knowledge that informs public decision-making processes by incorporating advances in machine learning and spatial analysis. (AU)