This repository contains the R programming scripts used for model development and implementation for the the paper entitled "A Bayesian statistical model-based approach for disaggregating small area population estimates by demographic characteristics". We present a flexible model-based appraoch for disaggregating small area population estimates by age, sex and social groups in order to facilitate more efficient target group-based interventions. The methodology is a multi-level Binomial regression model-based appraoch and parameter inference was based on Bayesian statisics, thus allowing easy uncertainty quantification.
There are two major R scripts contained in this repository
- Application_main_final.R: This includes the method application script which shows the step by step appraoch for the method implementation using datasets on Cameroon as a case study.
- simulation_study-main_final.R: The R scripts developed for the purpose of simulation study which aimed primarily to evaluate method performance over different sample sizes.
The methodology requires different layers of datasets from sanple survey data to geospatial covariates. The sample survey provides the population group structure of interest across some subpopulations. These are the number of people within the administrative unit belonging to each of the characteristics of interest (e.g., number of people aged '15 - 49' years old, number of people from the 'richer' households). Please note we have only included the simulated data in this repository. The application data may be obtained from the National Statistical Office, Cameroon.
To facilitate the use of the methodology described in the paper, we have developed a user-friendly R package known as "JollofR version 0.3.0" which automatically disaggregates population estimates at both administrative uni and grid cell levels, along with the estimates of uncertainty. Please visit https://github.com/wpgp/jollofR to learn more about the "jollofR" package.