Dengue mortality by DENV serotype in Mexico, 2021–2025: minimal reproducible dataset

This dataset provides the minimal, de-identified, complete-case cohort required to reproduce the tables, figures, and statistical models reported in "Serotype-specific dengue mortality in Mexico: a comparison of regression-based and machine learning approaches". Data derive from Mexico's national dengue surveillance system (Secretaría de Salud, General Directorate of Epidemiology), restricted to PCR-confirmed cases with an identified DENV serotype (DENV-1 to DENV-4), symptom onset between January 2021 and December 2025. The dataset includes age, sex, DENV serotype, season of symptom onset, dengue-related mortality, and eight clinical/demographic comorbidities (self-identified Indigenous status, hemorrhagic manifestations, diabetes, hypertension, peptic ulcer disease, chronic kidney disease, immunosuppression, and liver cirrhosis). All records are anonymized; no direct identifiers are included. Used together with the accompanying analysis pipeline (R, archived separately on GitHub), this dataset supports full reproduction of the manuscript's primary multivariable logistic regression, core-predictor comparison panel (standard logistic regression, GAM, LASSO, random forest), and all associated performance and stability analyses.