Raw Data, Compiled Database, and Reproducible Scripts: An Integrated Brazilian Repository on Occupational Accidents and Climate (2006–2024)

This repository presents a complete, academic resource detailing the data and code used to construct a rigorous Brazilian database linking occupational accident records from the National Notifiable Diseases Information System (SINAN) with comprehensive meteorological variables collected by the National Meteorology Institute (INMET). The resulting linked dataset covers an 18-year period, from 2006 to 2024, and is structured to serve as a verifiable foundation for climate-health research across Brazil. The repository follows a clear, multi-stage data workflow to ensure maximum transparency and reproducibility. It contains the original, unprocessed raw data, including municipal geographic coordinates (IBGE), SINAN accident records, and INMET weather station data (both daily and hourly readings). This is followed by the imputed datasets, where missing meteorological observations were systematically addressed using the Amelia package in R to preserve the spatio-temporal structure of the variables. The final analytical product is the compiled database, which integrates the two sources. This linking relies on a critical spatial-temporal matching process: the Haversine formula was employed to calculate distances and assign meteorological data from the closest available station to each accident record, strictly adhering to a maximum proximity criterion of 150 kilometers. Crucially, all necessary, fully commented R scripts used for data preparation, imputation, and the final database compilation are included. This structure ensures that the entire dataset creation process is auditable and fully reproducible, making the repository an invaluable resource for empirical analysis into how climatic variations affect occupational health outcomes.