babamaiyaki200016-prog/new-diphtheria-nigeria

# Nigeria Diphtheria Outbreak: Kano Line-List Sub-Analysis R scripts for the Kano case-level sub-analysis supplementing "Epidemiology, Mortality, and Intervention Strategies for Diphtheria Control: A Comprehensive Analysis of Nigeria's 2022-2026 Outbreak" (Musa BM et al.). Rebuilds the outbreak analysis from a 22,379-row Kano line list (2022-05 to 2024-05), replacing several sitrep-based estimates in the manuscript with case-level analysis: a real (not back-calculated) epidemic curve, direct + Bayesian hierarchical age-specific case fatality rates, a stable EpiEstim Rt estimate, and a Bayesian SEIR transmission model (custom Metropolis-Hastings MCMC in base R -- no Stan/rstanarm dependency required to rerun these scripts). ## Running Scripts run in order from the project root (working directory containing `data/`, `figures/`, `scripts/`): ``` Rscript scripts/00_setup.R Rscript scripts/01_clean_linelist.R Rscript scripts/02_epidemic_curve.R Rscript scripts/03_age_cfr.R Rscript scripts/04_rt_estimation.R Rscript scripts/05_bayesian_transmission_model.R Rscript scripts/06_vaccination_scenarios.R Rscript scripts/07_policy_impact.R Rscript scripts/08_geographic_distribution.R ``` Each script reads its inputs from `data/` and writes intermediate CSVs back to `data/` and figures to `figures/`; run them in order the first time. Input data (the raw Kano line list, and small reference CSVs transcribed from the manuscript's own tables) are not included in this repository. ## Methodology notes - `05_bayesian_transmission_model.R`'s header comment explains the SEIR model structure, why it's fit to a 27-week window rather than the full 91-week series, and how the effective population parameter (N_eff) is a fitted scale, not Kano's real population. - Bayesian components use custom Metropolis-Hastings sampling rather than Stan/rstanarm, for portability across environments without a working C++/Stan toolchain.