Hotelling's T² Ellipses and Outlier Detection for 'ggplot2'

Calculate Hotelling's T² ellipses and detect multivariate outliers both for base R plots and 'ggplot2' plots. Optionally, uses robust covariance estimation to reduce the influence of outliers on the ellipses. Also included: bagplots, kernel density plots and outlier diagnostic plots.

Hotelling’s T² Ellipses and Outlier Detection for ggplot2

Here are a few examples of plots you can create with gghotelling:

Features:

• Classical Hotelling and data ellipses with geom_hotelling()
• Robust Hotelling ellipses using MCD estimator with robust=TRUE
• Hotelling confidence ellipses for group means with type="t2mean"
• Kernel density coverage contours with geom_kde()
• Outlier detection with outliers()
• Outlier visualization with stat_outliers() and plot_outliers()
• Convex hulls with geom_hull()
• Bagplots with geom_bag()
• Autoplot and autolayer methods for prcomp objects

See the vignette for more details.

Installation

You can install the development version of gghotelling from GitHub with:

# install.packages("pak")
pak::pak("january3/gghotelling")

Minimal Example

The package defines a new geom, geom_hotelling(), which can be used to add Hotelling ellipses to ggplot2 scatter plots. Here are two examples of usage:

library(ggplot2)
library(gghotelling)

pca <- prcomp(iris[, 1:4], scale.=TRUE)
df <- cbind(iris, pca$x)

# minimum example
p1 <- ggplot(df, aes(PC1, PC2, color=Species)) +
  geom_hotelling() +
  geom_point()

# set custom CI/coverage level
p2 <- ggplot(df, aes(PC1, PC2, color=Species)) +
  geom_hotelling(alpha=0.1, aes(fill = Species), level=.99) +
  geom_point()

cowplot::plot_grid(p1, p2, ncol=2)

AI and other tools disclosure

I used LLM modes to a) detect bugs, spelling errors, clerical mistakes and other deficiencies in the code and b) better understand the ggplot2 specific code (especially after_stat()).

Furthermore, I used devtools, checktor and covr.