ggtriangles provides the geom_triangles() function which plots isosceles triangles with their horizontal base centered at position (x,y) and variable triangle_height, triangle_width, and angle as given by aes() arguments in the ggplot2 framework.

note: ggtriangles is still under development.

Here, the triangle_height argument is used to show the hypothesized effect of the drug (see details in ?datasets::sleep) given on each individual. ggtriangles supports the automatic creation of the legend shown here -- see how to make it in the vignette('intro-to-ggtriangles').

library(ggtriangles)
data.frame(x = seq(0,1,.1)) %>%
  ggplot(aes(x = x, y = x, triangle_height = x)) +
  geom_triangles()

The geom_triangles() layer takes aesthetic arguments for triangle_height, triangle_width, and angle that can be mapped to data. Here, the x-axis, y-axis, and triangle_height visual channels are all mapped to the same set of data.

devtools::install_github("ctesta01/ggtriangles")

Read about how to make these in the vignette('intro-to-ggtriangles').

Check out the examples available in the help pages by entering ?geom_triangles into R after you've loaded the package. Additionally, vignette('intro-to-ggtriangles') has the code to produce the figures above.

I think ggtriangles will provide a nice interface both for Rtists who are looking to incorporate triangles programmatically in their generative art, as well as data visualization creators who are interested in using triangles' position, height, width, color, and fill to communicate features of their data.

I originally created this because I wanted to be able to render something like the NYTimes maps of where COVID-19 death rates had increased and decreased since vaccines were widely made available to adults.

These maps were originally featured here: https://www.nytimes.com/interactive/2021/12/28/us/covid-deaths.html

I spent some time re-creating something similar (though not exactly the same) as the NYTimes analysis and graphic to demonstrate that functionality and use-case for ggtriangles.

You can check out how I did it in the vignette available through the vignette available in browseVignettes('ggtriangles') or the source code, vignettes/recreating_nytimes_viz.Rmd. Note that these figures are provided purely as examples and not meant for decision making purposes or scientific usage. I'm pleased with how well I was able to replicate their plot of where death rates had increased, but I can't say for sure why I appear to be getting much larger decreases in the COVID-19 death rates per 100,000 than they did in their analysis.

Please feel free to make suggestions and provide feedback on ggtriangles in an issue on this GitHub repository.