This repository contains code, data, and figure/table files for an analysis of reproductive effects on bat viral detection that leverages data from the USAID-funded PREDICT project. More specifically, data for this project were accessed from the Emering Infectious Disease Information Technology Hub (EIDITH) database using the eidith R package.

Repository contents

The contents of the repository subdirectories are as follows:

• R/ contains internal functions used in the repository code

• data/ contains various lookup tables used during data cleaning as well as the cleaned data for model-fitting

• outputs/ contains all figures and tables associated with the analysis

• scripts/ contains the core repository analyses:

  • 01_data_prep.R accesses, cleans, and filters EIDITH data for further analysis
  • 02_stan_model_fitting.R fits hierarchical Bayesian models to the cleaned datasets using rstan
  • 03_plotting.R generates all figures

• stan/ contains cleaned data subsets in Stan-friendly formats and the Stan models used in the analysis

Reproducing the analysis

In order to facilitate reproducibility for those without EIDITH access, cleaned data analyzed here are contained in both the data/cleaned_data subdirectory (full analysis set in CSV format) and the stan/cleaned_data subdirectory (full analysis set and data subsets as Stan-friendly list objects). Users cloning the repository will therefore be able to proceed with the recreation of the analysis starting with scripts/02_stan_model_fitting.R, which will fit the Stan models, saving them to stan/saved_models. Subsequently, 03_plotting.R will recreate all figures using this saved model output.