RiskMap

Overview

RiskMap provides tools for model-based geostatistical analysis of continuous, binomial and Poisson outcomes.

• Fit spatial and spatio-temporal Gaussian process models.
• Generate predictive surfaces and target summaries.
• Run simulation-based diagnostics and validation workflows.

The methodology is described in Model-based Geostatistics for Global Public Health by Diggle and Giorgi.

Start Here: MBG-R Book

For a full applied guide to using RiskMap in real public health workflows, see the online book by Emanuele Giorgi and Claudio Fronterre:

Model-based geostatistics for global public health using R https://www.mbgr.org/

Installation

Install the stable version from CRAN:

install.packages("RiskMap")

Install the development version from GitHub:

# install.packages("devtools")
devtools::install_github("claudiofronterre/RiskMap")

Quickstart

A minimal linear Gaussian geostatistical model of the form $Y(x) = \beta_0 + S(x)$, where $S(x)$ is a spatial Gaussian process can be fitted with:

library(RiskMap)

data(italy_sim)

fit <- glgpm(
  formula = y ~ gp(x1, x2),
  data = italy_sim,
  family = "gaussian",
  crs = 32634,
  messages = FALSE
)

summary(fit)

Learn More

• MBG-R book: https://www.mbgr.org/
• Package website: https://claudiofronterre.github.io/RiskMap/
• Issue tracker: https://github.com/claudiofronterre/RiskMap/issues