Code and data for: Cold days, access to stroke centers, and stroke admissions, deaths, and case fatality in southern Brazil, 2012–2024

Code and processed data supporting the article "Cold days are associated with more stroke admissions and deaths, and distance to stroke centers with higher in-hospital case fatality: time series in southern Brazil, 2012–2024". The study covers the 69 health regions (1,191 municipalities) of Paraná, Santa Catarina, and Rio Grande do Sul, Brazil, from 2012 to 2024. It links daily temperature (ERA5-Land reanalysis) to 147,634 stroke deaths from the Mortality Information System (SIM) and 496,166 stroke admissions from the Hospital Information System of the Brazilian Unified Health System (SIH/SUS), using a case time series design with conditional quasi-Poisson regression and distributed lag non-linear models (lag 0–21 days). It also relates road travel time to the nearest accredited stroke center (CNES, OSRM on OpenStreetMap) to in-hospital case fatality. Contents: Python and R scripts for every step (data preparation, models, figures, and tables), processed input data (aggregated, with no personal identifiers), the extraction log and data dictionary, and the resulting tables and figures. Running the pipeline reproduces all tables and figures in the article. See README.md for the folder structure, data sources, and execution order. Code is released under the MIT License and processed data under CC BY 4.0, subject to the terms of the original sources (DATASUS, Copernicus/ERA5, Open-Meteo, OpenStreetMap, and IBGE).