Spatiotemporal Dataset and Reproducible Pipeline for Adolescent Suicide Mortality in Colombia (2015–2024)
Description
This dataset contains the supplementary materials, analytical datasets, code, model outputs, and documentation supporting the manuscript “Spatiotemporal patterns and geographic inequality in registered adolescent suicide mortality in Colombia, 2015–2024: a Bayesian disease-mapping study”. The materials provide a reproducible analytical pipeline for examining registered adolescent suicide mortality among individuals aged 10–19 years in Colombia between 2015 and 2024. The analysis is based on a balanced department-year panel integrating mortality counts, population denominators, socioeconomic covariates, and spatial adjacency structures. The repository includes harmonized and derived data products from publicly available sources, including official medico-legal mortality records from the Instituto Nacional de Medicina Legal y Ciencias Forenses (INMLCF), population denominators from the Departamento Administrativo Nacional de Estadística (DANE), Multidimensional Poverty Index (MPI) covariates, and official Colombian administrative boundary data. The analytical workflow is organized into five modules: 00_population: population denominators derived from DANE historical and projected population series. 01_MPI: processing and harmonization of Multidimensional Poverty Index covariates. 02_panel: construction of the balanced analytical panel and descriptive summaries. 03_Spatial: construction of spatial adjacency structures and graph diagnostics. 04_bayes: Bayesian hierarchical modeling scripts, posterior summaries, model diagnostics, sensitivity analyses, and supplementary tables. Final manuscript-ready supplementary tables are provided in 04_bayes/build/, including Tables S1–S21 and compiled numerical outputs. No individual-level identifiable data are included in the shared materials. Intermediate files are retained to support transparency, traceability, and reproducibility.
Files
Steps to reproduce
STEPS TO REPRODUCE To reproduce the analytical workflow, run the scripts sequentially in the order below. 1. Prepare population denominators Run: 00_population/00_prepare_population_lookup.R 2. Harmonize socioeconomic covariates Run: 01_MPI/01_prepare_MPI_covariates.R 3. Construct the analytical panel Run: 02_panel/02_prepare_suicide_panel.R 4. Build the spatial adjacency structure Run: 03_Spatial/03_prepare_spatial_graph_DP33.R 5. Run Bayesian spatiotemporal models Run: 04_bayes/04_Bayesian_Model.R This script fits the main Bayesian hierarchical models and exports posterior summaries, model diagnostics, stratified estimates, and core model outputs. 6. Generate post hoc model metrics Run: 04_bayes/04b_posthoc_metrics.R This script derives additional post hoc quantities from the Bayesian model outputs, including secondary summaries, comparisons, and diagnostic metrics used in the supplementary results. 7. Run hyperprior sensitivity analyses Run: 04_bayes/04c_hyperprior_sensitivity.R This script evaluates the sensitivity of the Bayesian estimates to alternative prior or hyperprior specifications and exports the corresponding robustness outputs. 8. Generate supplementary tables Run: 04_bayes/05_supplementary_tables.R This script generates the final supplementary tables and compiled numerical outputs, including Tables S1–S21, stored in: 04_bayes/build/