README.md

In brianconroy/preferential_surveillance: Analysis of Disease Surveillance Data under Preferential Sampling

Preferential Surveillance

This repository contains code to estimate disease risk surfaces from preferentially sampled disease surveillance data. It includes simulation studies comparing performance of the proposed method against various benchmarks (Bayesian Additive Regression Trees, spatial Poisson regression, Poisson regression). Also included is an analysis of real surveillance data monitoring plague (Yersinia pestis) within the rodent population of California.

This repository can be installed like a package for ease of access to function documentation, processed datasets, and dependency management. However it also contains a number of "project like" elements including simulation and analysis scripts.

Simulation scripts are organized as follows:

• simulation_data.R: generates simulated disease surveillance datasets under differing levels of preferential sampling.
• simulation.R: Fits the proposed preferential sampling method along with benchmarks to the datasets simulated by simulation_data.R.
• simulation_analysis.R: Compares performance of the models fit by simulation.R.

Plague analysis scripts are structured as:

• data_analysis_fits.R: Fits the proposed preferential sampling method along with benchmarks to the processed plague surveillance dataset. Additionally, performs spatial downscaling to obtain high resolution risk maps.
• data_analysis_summary.R: Summarizes, visualizes, and compares risk maps obtained from the proposed method and benchmarks over theplague surveillance dataset.

Remaining files are organized as in any other package:

• R/: holds all helper functions
• data-raw: contains scripts to generate the processed datasets of the package
• man/: documentation

Installation

Assumptions:

• R version 3.4.3 - 4.0.5 has been installed.
• The devtools package has been installed.

install.packages("devtools")

Dependencies of the repository and proccessed datasets can be installed and made accessible by first cloning the repository, navigating to the root of the repository in the R console, and installing the repository locally via:

# install package dependencies
devtools::install_deps()

# install package
devtools::install()

Successful installation can be checked by loading the package and inspecting its datasets.

library(preferentialSurveillance)

# first and second principal components of the PRISM climatic dataset
plot(prism_pca)

Data Access

Caveat. The scripts for data analysis in this repo consume a plague surveillance dataset of aggregate (1983-2015) counts of plague positive and negative rodents. This dataset was collected and is maintained by the California Department of Public Health - Vector Borne Disease Section. Due to department policy the data are not included in the repository, but are available upon request.

brianconroy/preferential_surveillance documentation built on Nov. 23, 2021, 5:51 a.m.