I have an issue where I would like to calculate locations of y-axis labels in a large plot mad with facet_grid()
. Let me show you what I mean using the mpg
dataset.
require(tidyverse)
ggplot(mpg, aes(x = displ)) +
geom_point(aes(y = hwy)) +
facet_grid(drv ~ class, scales = "free")
You will notice that both axes use variable numbers of labels. Ignoring problems with the x-axis, I am interested in labelling only three values on the y-axis: 0, half of max value, and max value. It makes sense in my use-case, so here is what I tried.
ggplot(mpg, aes(x = displ)) +
geom_point(aes(y = hwy)) +
facet_grid(drv ~ class, scales = "free") +
geom_blank(aes(y = 0)) + # extends y-axis to 0
scale_y_continuous(expand = expansion(mult = c(0, 0.1)), # prevents ggplot2 from extending beyond y = 0
n.breaks = 3) # Three axis labels, please.
The plot correctly starts at y = 0 and labels it correctly. However, remaining labels are poorly assigned, and have labels n = 2
and n = 4
instead of n = 3
for some reason.
If I could only directly calculate the label positions!
ggplot(mpg, aes(x = displ)) +
geom_point(aes(y = hwy)) +
facet_grid(drv ~ class, scales = "free") +
geom_blank(aes(y = 0)) +
scale_y_continuous(expand = expansion(mult = c(0, 0.1)),
n.breaks = 3,
breaks = c(0, 0.5*max(hwy), 1*max(hwy))) # Maybe these formulas works?
Error in check_breaks_labels(breaks, labels) : object 'hwy' not found
I believe providing break points by this way should work, but that my syntax is bad. How do I access and work with the data underlying the plot? Alternatively, if this doesn't work, can I manually specify y-axis labels for each row of panels?
I could really use some assistance here, please.
If you want custom rules for breaks, the easiest thing is to use a function implementing those rules given the (panel) limits.
Below an example for labeling 0, the max and half-max.
library(ggplot2)
ggplot(mpg, aes(x = displ)) +
geom_point(aes(y = hwy)) +
facet_grid(drv ~ class, scales = "free") +
scale_y_continuous(expand = expansion(mult = c(0, 0.1)),
limits = c(0, NA), # <- fixed min, flexible max. replaces geom_blank
breaks = function(x){c(x[1], mean(x), x[2])})