Fig. 4

Maps of unstructured and structured random effects of new HIV diagnosis, Siaya County. a) Map of estimated median value of unstructured random effects, showing residual variability of new HIV diagnoses when spatial autocorrelation was not taken into account; b) Map of estimated median value of structured random effects, showing residual variability of new HIV diagnoses when spatial autocorrelation was accounted for. The maps were generated from the convolution Bayesian Poisson model, and mapped using ggplot2 R package (https://ggplot2.tidyverse.org)