How to make a base R style boxplot using ggplot2?

Question

I need to make a lot of boxplots for an upcoming publication. I would like to use ggplot2 because I think it will be more flexible for future projects, but my PI is insisting that I make these plots in the style of base-R. He specifically wants the dashed lines, so that they will appear similar to previous plots we made. I have made an example using the iris dataset to show you, using this code:

plot(iris$Species,
     iris$Sepal.Length,
     xlab='Species',
     ylab='Sepal Length',
     main='Sepal Variation Across Species',
     col='white')

My question is how to make a similar looking plot using ggplot2?

Here is my attempt:

library("ggplot2")
ggplot(iris) +
  geom_boxplot(aes(x=Species,y=Sepal.Length),linetype="dashed") +
  ggtitle("Sepal Variation Across Species")

I need the combination of dashed and solid lines, but I cannot make anything work. I have already checked https://stats.stackexchange.com/questions/8137/how-to-add-horizontal-lines-to-ggplot2-boxplot which is very very close but no dashed lines, which we need. Also the outliers are filled circles, which is not the same as base-R.

Smedley

Use outlier.color = 'black' and outlier.fill = 'white' to reproduce the circles. – Hodgkins

Unfortunately there are no individual/separate aesthetic mappings for the IQR or median line but you could make your own version of geom_boxplot() and modify what gets passed into github.com/tidyverse/ggplot2/blob/master/R/… by adding such params. – Classical

Answer 1

To generate a "base R style" boxplot using ggplot2, we can layer 4 boxplot objects over top of one another. The order does matter here, so please keep this in mind if you modify the code. I strongly suggest that you explore this code by plotting each boxplot layer on its own; that way you can get a feel for how the different layers interact.

The ordering of the boxplots works like this (ordered from bottom to top):

• (1) vertical dashed lines are placed first
• (2) a solid box containing a median line, which covers the dashed box from (1)
• (3) & (4) solid whisker lines, created by using errorbars with the minima set to the maxima, and vice versa.

I also added custom breaks to match your base R plot, which you can change depending on your needs. panel.border is used to create a thin border in the style of base R. To get the open circles that you want, we use outlier.shape.

The code:

library("ggplot2")

ggplot(data = iris, aes(x = Species, y = Sepal.Length)) +
  geom_boxplot(linetype = "dashed", outlier.shape = 1) +
  stat_boxplot(aes(ymin = ..lower.., ymax = ..upper..), outlier.shape = 1) +
  stat_boxplot(geom = "errorbar", aes(ymin = ..ymax..)) +
  stat_boxplot(geom = "errorbar", aes(ymax = ..ymin..)) +
  scale_y_continuous(breaks = seq(4.5, 8.0, 0.5)) +
  labs(title = "Sepal Variation Across Species",
       x = "Species",
       y = "Sepal Length") +
  theme_classic() + # remove panel background and gridlines
  theme(plot.title = element_text(hjust = 0.5,  # hjust = 0.5 centers the title
                                  size = 14,
                                  face = "bold"),
        panel.border = element_rect(linetype = "solid",
                                    colour = "black", fill = "NA", size = 0.5))

The plot:

Not quite exactly the same, but it seems to be a decent approximation. Hopefully this is close enough for your needs. Good luck, and happy plotting!

Offish

Answer 2

Here's a wrapper around @Marcus' great solution, for convenient use and more flexibility:

geom_boxplot2 <- function(mapping = NULL, data = NULL, stat = "boxplot", position = "dodge2", 
                          ..., outlier.colour = NULL, outlier.color = NULL, outlier.fill = NULL, 
                          outlier.shape = 1, outlier.size = 1.5, outlier.stroke = 0.5, 
                          outlier.alpha = NULL, notch = FALSE, notchwidth = 0.5, varwidth = FALSE, 
                          na.rm = FALSE, show.legend = NA, inherit.aes = TRUE,
                          linetype = "dashed"){
  list(
    geom_boxplot(mapping = mapping, data = data, stat = stat, position = position,
                 outlier.colour = outlier.colour, outlier.color = outlier.color, 
                 outlier.fill = outlier.fill, outlier.shape = outlier.shape, 
                 outlier.size = outlier.size, outlier.stroke = outlier.stroke, 
                 outlier.alpha = outlier.alpha, notch = notch, 
                 notchwidth = notchwidth, varwidth = varwidth, na.rm = na.rm, 
                 show.legend = show.legend, inherit.aes = inherit.aes, 
                 linetype = linetype, ...),
    stat_boxplot(aes(ymin = ..lower.., ymax = ..upper..), outlier.shape = 1) ,
    stat_boxplot(geom = "errorbar", aes(ymin = ..ymax..)) ,
    stat_boxplot(geom = "errorbar", aes(ymax = ..ymin..)) ,
    theme_classic(), # remove panel background and gridlines
    theme(plot.title = element_text(hjust = 0.5,  # hjust = 0.5 centers the title
                                    size = 14,
                                    face = "bold"),
          panel.border = element_rect(linetype = "solid",
                                      colour = "black", fill = "NA", size = 0.5))
  )
}

ggplot(data = iris, aes(x = Species, y = Sepal.Length)) +
  geom_boxplot2() +
  scale_y_continuous(breaks = seq(4.5, 8.0, 0.5)) + # not sure how to generalize this
  labs(title = "Sepal Variation Across Species", y = "Sepal Length")

Minnesinger

Answer 3

Building further on what @Marcus & @Moody_Mudskipper has provided:

geom_boxplotMod <- function(mapping = NULL, data = NULL, stat = "boxplot", 
    position = "dodge2", ..., outlier.colour = NULL, outlier.color = NULL, 
    outlier.fill = NULL, outlier.shape = 1, outlier.size = 1.5, 
    outlier.stroke = 0.5, outlier.alpha = NULL, notch = FALSE, notchwidth = 0.5,
    varwidth = FALSE, na.rm = FALSE, show.legend = NA, inherit.aes = TRUE,
    linetype = "dashed") # to know how these come here use: args(geom_boxplot)
    {
    list(geom_boxplot(
            mapping = mapping, data = data, stat = stat, position = position,
            outlier.colour = outlier.colour, outlier.color = outlier.color, 
            outlier.fill = outlier.fill, outlier.shape = outlier.shape, 
            outlier.size = outlier.size, outlier.stroke = outlier.stroke, 
            outlier.alpha = outlier.alpha, notch = notch, 
            notchwidth = notchwidth, varwidth = varwidth, na.rm = na.rm, 
            show.legend = show.legend, inherit.aes = inherit.aes, linetype = 
            linetype, ...),
        stat_boxplot(geom = "errorbar", aes(ymin = ..ymax..), width = 0.25),
        #the width of the error-bar heads are decreased
        stat_boxplot(geom = "errorbar", aes(ymax = ..ymin..), width = 0.25),
        stat_boxplot(aes(ymin = ..lower.., ymax = ..upper..),
            outlier.shape = 1),
        theme(panel.background = element_blank(),
            panel.border = element_rect(size = 1.5, fill = NA),
            plot.title = element_text(hjust = 0.5),
            axis.title = element_text(size = 12),
            axis.text = element_text(size = 10.5))
        )
    }

library(tidyverse); library(ggplot2);
ggplot(iris, aes(x=Species,y=Sepal.Length, colour = Species)) +
    geom_boxplotMod() +
    ggtitle("Sepal Variation Across Species")

^{Created on 2020-07-20 by the reprex package (v0.3.0)}

Topeka