SaviKoissi/BBTD_Transmission_Mechanism

SAVI, M. K. (2026). Pathways of Banana Bunchy Top Disease in Benin: Structural Canalization. Zenodo. doi.org # Tracing the Pathways of Banana Bunchy Top Disease in Benin ## 1. Overview This project investigates the dominant transmission mechanisms of Banana Bunchy Top Disease (BBTD) in Benin using a spatially explicit stochastic metapopulation framework. The analysis is built as a modular pipeline that integrates: * Field-level observational data * Bias-adjusted prevalence estimation * Spatial data processing * Bayesian inference via Stan * Simulation-based inference (SMC-ABC) for transmission dynamics The goal is to infer population-level transmission processes, not individual infection histories. ## 2. Scope of This Repository This repository contains the data processing and observational inference layer of the broader modeling framework. It represents the data-driven foundation upon which the simulation and ABC-SMC inference pipeline is built. ⚠️ The simulation and model calibration components (IBM, SMC-ABC) are implemented in a separate computational workflow and are not fully contained in this repository. ## 3. Project Philosophy The workflow is designed to be: * Modular → each stage is independent and reusable * Reproducible → all steps are script-based and traceable * Scalable → compatible with high-performance computing workflows * Statistically rigorous → integrates Bayesian inference with uncertainty quantification ## 4. Data Pipeline Overview The pipeline processes raw survey and spatial data into structured inputs suitable for downstream modeling. * Core Steps * Data Cleaning * Standardization of field identifiers * Handling missing values * Harmonization of survey formats * Historical Data Integration * Incorporation of archival surveillance records * Alignment of temporal references * Spatial Processing * Mapping farms to administrative boundaries * Linking geospatial shapefiles with field data * Prevalence Estimation * Bayesian estimation of field-level prevalence * Adjustment for diagnostic misclassification (sensitivity/specificity) * Data …