Chernobyl Fallout Proxies and Provincial Cancer Mortality in Türkiye (2018–2024): Reproducible Data and Analysis Code

Summary. This record provides the data and reproducible Python code used to analyze associations between spatial proxies of the Chernobyl accident (great-circle distance; province-level ^137Cs deposition) and provincial neoplasm mortality in Türkiye (2018–2024). We include PCA correlation circle outputs, penalized regression (Ridge/Lasso/Elastic-Net) with repeated holdout and nested CV, bootstrap coefficient intervals, and spatial diagnostics (global Moran’s I, LISA). Contents.– DATA.xlsx (analysis-ready dataset; variable keys in DATA_DICTIONARY.md)– CITY_NEIGHBORS.xlsx (adjacency list)– build_residuals.py, moran_analysis.py, pca_correlation_circle.py, penalized_regression_pipeline.py– requirements.txt, environment.yml, LICENSE, README.md, CHANGELOG.md, checksums.txt How to reproduce. See README.md for environment setup and step-by-step commands (PCA, penalized models, and spatial diagnostics producing Figure S1 panels). Licensing. Dataset and example figures: CC BY 4.0. Source code: MIT (see LICENSE). Related publication. [Article title], Environmental Research, [journal DOI if available]. Keywords. Chernobyl; neoplasm mortality; Türkiye; spatial epidemiology; Moran’s I; LISA; penalized regression; PCA.