An R package for visualizing uncertainty in spatial data.

Vizumap pkgdown site with vignette: https://lydialucchesi.github.io/Vizumap/.

Installation

You can install a development version of the Vizumap package using the command below.

remotes::install_github(repo = "lydialucchesi/Vizumap", build_vignettes = TRUE, force = TRUE)

Authors

Lydia Lucchesi, Australian National University & CSIRO Data61, Email:

Petra Kuhnert, CSIRO Data61, Email:

About the package

Approaches for visualising uncertainty in spatial data are presented in this package. These include the three approaches developed in Lucchesi and Wikle (2017) and a fourth approach presented in Kuhnert et al. (2018).

Bivariate Maps

In these bivariate choropleth maps, two colour schemes, one representing the estimates and another representing the margins of error, are blended so that an estimate and its error can be conveyed on a map using a single colour.

Map Pixelation

In this approach, each map region is pixelated. Pixels are filled with colours representing values within an estimate’s margin of error. Regions that appear as a solid colour reflect smaller margins of error, while more pixelated regions indicate greater uncertainty. These maps can be animated to provide a novel uncertainty visualisation experience.

Glyph Rotation

In this method, glyphs located at region centroids are rotated to represent uncertainty. The colour filling each glyph corresponds to the estimate.

Exceedance Probability Maps

The final map-based exploration is through exceedance probabilities, which are visualised on a map to highlight regions that exhibit varying levels of departure from a threshold of concern or target.

Examples

A for the Vizumap package is available and contains examples relating to each of the visualisation methods.

vignette("Vizumap")

Testing

If you would like to install and run the unit tests interactively, include INSTALL_opts = "--install-tests" in the installation code.

remotes::install_github(repo = "lydialucchesi/Vizumap", build_vignettes = TRUE, force = TRUE, INSTALL_opts = "--install-tests")

testthat::test_package("Vizumap", reporter = "stop")

Contribute

To contribute to Vizumap, please follow these guidelines.

Please note that the Vizumap project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

License

Vizumap version 1.2.0 is licensed under GPLv3.

History of Vizumap

Vizumap began as a visualisation project at the University of Missouri in 2016. Chris Wikle, professor of statistics, posed an interesting research question to Lydia Lucchesi, a student curious about data visualisation and R.

How do you include uncertainty on a map displaying areal data estimates?

Over the course of a year, they put together three methods for visualising uncertainty in spatial statistics: the bivariate choropleth map, the pixel map, and the glyph map. By mid-2017, there were maps, and there was a lot of R code, but there was not a tool that others could use to easily make these types of maps, too. That’s when statistician Petra Kuhnert recommended developing an R package. Over the course of a month, Petra and Lydia developed Vizumap (originally named VizU) at CSIRO Data61 in Canberra, Australia. Since then, the package has been expanded to include exceedance probability maps, an uncertainty visualisation method developed by Petra while working on a Great Barrier Reef (GBR) project.

Vizumap has been used to visualise the uncertainty of American Community Survey estimates, the prediction errors of sediment estimates in a GBR catchment, and most recently the uncertainty of estimated locust densities in Australia. We would like to assemble a Vizumap gallery that showcases different applications of the package’s mapping methods. If you use Vizumap to visualise uncertainty, please feel free to send the map our way. We would like to see it!