Combination Boxplot and Histogram using ggplot2

Question

I am trying to combine a histogram and boxplot for visualizing a continuous variable. Here is the code I have so far

require(ggplot2)
require(gridExtra)
p1 = qplot(x = 1, y = mpg, data = mtcars, xlab = "", geom = 'boxplot') + 
     coord_flip()
p2 = qplot(x = mpg, data = mtcars, geom = 'histogram')
grid.arrange(p2, p1, widths = c(1, 2))

It looks fine except for the alignment of the x axes. Can anyone tell me how I can align them? Alternately, if someone has a better way of making this graph using ggplot2, that would be appreciated as well.

Goldstein

Your original question was about how to achieve using ggplot, however, the answer you have marked 'accepted' utilized qplot. Which is a different thing. Which may serve the purposed, though, we can see that now there is a ggplot answer below. – Tjader

Answer 1

you can do that by coord_cartesian() and align.plots in ggExtra.

library(ggplot2)
library(ggExtra) # from R-forge

p1 <- qplot(x = 1, y = mpg, data = mtcars, xlab = "", geom = 'boxplot') + 
  coord_flip(ylim=c(10,35), wise=TRUE)
p2 <- qplot(x = mpg, data = mtcars, geom = 'histogram') + 
  coord_cartesian(xlim=c(10,35), wise=TRUE)

align.plots(p1, p2)

Here is a modified version of align.plot to specify the relative size of each panel:

align.plots2 <- function (..., vertical = TRUE, pos = NULL) 
{
    dots <- list(...)
    if (is.null(pos)) pos <- lapply(seq(dots), I)
    dots <- lapply(dots, ggplotGrob)
    ytitles <- lapply(dots, function(.g) editGrob(getGrob(.g, 
        "axis.title.y.text", grep = TRUE), vp = NULL))
    ylabels <- lapply(dots, function(.g) editGrob(getGrob(.g, 
        "axis.text.y.text", grep = TRUE), vp = NULL))
    legends <- lapply(dots, function(.g) if (!is.null(.g$children$legends)) 
        editGrob(.g$children$legends, vp = NULL)
    else ggplot2:::.zeroGrob)
    gl <- grid.layout(nrow = do.call(max,pos))
    vp <- viewport(layout = gl)
    pushViewport(vp)
    widths.left <- mapply(`+`, e1 = lapply(ytitles, grobWidth), 
        e2 = lapply(ylabels, grobWidth), SIMPLIFY = F)
    widths.right <- lapply(legends, function(g) grobWidth(g) + 
        if (is.zero(g)) 
            unit(0, "lines")
        else unit(0.5, "lines"))
    widths.left.max <- max(do.call(unit.c, widths.left))
    widths.right.max <- max(do.call(unit.c, widths.right))
    for (ii in seq_along(dots)) {
        pushViewport(viewport(layout.pos.row = pos[[ii]]))
        pushViewport(viewport(x = unit(0, "npc") + widths.left.max - 
            widths.left[[ii]], width = unit(1, "npc") - widths.left.max + 
            widths.left[[ii]] - widths.right.max + widths.right[[ii]], 
            just = "left"))
        grid.draw(dots[[ii]])
        upViewport(2)
    }
}

usage:

# 5 rows, with 1 for p1 and 2-5 for p2
align.plots2(p1, p2, pos=list(1,2:5))
# 5 rows, with 1-2 for p1 and 3-5 for p2
align.plots2(p1, p2, pos=list(1:2,3:5))

Magill

Thanks Koshke!! I did try align.plots before, but without the coord_cartesian option, and it produced the same result as grid.arrange. Is there any way to specify the heights of the two plots as in grid.arrange? I would like the boxplot to be smaller in height than the histogram. – Goldstein

Thanks again Kohske. That works perfectly. Maybe you should send this patch to the authors of ggExtra, as I believe it is an extremely useful feature!! – Goldstein

Answer 2

Using cowplot package.

library(cowplot)

#adding xlim and ylim to align axis.
p1 = qplot(x = 1, y = mpg, data = mtcars, xlab = "", geom = 'boxplot') + 
  coord_flip() +
  ylim(min(mtcars$mpg),max(mtcars$mpg))

p2 = qplot(x = mpg, data = mtcars, geom = 'histogram')+
  xlim(min(mtcars$mpg),max(mtcars$mpg))

#result
plot_grid(p1, p2, labels = c("A", "B"), align = "v",ncol = 1)

Thrashing

Answer 3

Another possible solution using ggplot2, however, so far I do not know how to scale the two plots in height:

require(ggplot2)
require(grid)

fig1 <- ggplot(data = mtcars, aes(x = 1, y = mpg)) +
  geom_boxplot( ) +
  coord_flip() +
  scale_y_continuous(expand = c(0,0), limit = c(10, 35))

fig2 <- ggplot(data = mtcars, aes(x = mpg)) +
  geom_histogram(binwidth = 1) +
  scale_x_continuous(expand = c(0,0), limit = c(10, 35))

grid.draw(rbind(ggplotGrob(fig1),
                ggplotGrob(fig2),
                size = "first"))

Chaffin

Answer 4

The best solution I know is using the ggpubr package:

require(ggplot2)
require(ggpubr)
p1 = qplot(x = 1, y = mpg, data = mtcars, xlab = "", geom = 'boxplot') + 
     coord_flip()
p2 = qplot(x = mpg, data = mtcars, geom = 'histogram')
ggarrange(p2, p1, heights = c(2, 1), align = "hv", ncol = 1, nrow = 2)

Pilotage