Unit-independent position_nudge - r

Question

I would like to apply a position_nudge to an object, but it should always be a certain distance (e.g. in "cm") rather than relative to the scale of the measured variable.
data <- data.frame(
name=c("de","gb","cn","ir","ru") ,
value=c(3,12,5,18,45)*1
)
ggplot(data,
aes(x=name, y=value)) +
geom_bar(stat = "identity") +
geom_text(aes(y = 0,
label = paste0(name,value)),
position = position_nudge(y = -12)) +
coord_cartesian(ylim = c(0, 50), # This focuses the x-axis on the range of interest
clip = 'off') + # This keeps the labels from disappearing
theme(plot.margin = unit(c(1,1,1,1), "lines"))
When changing the scale of the variable, that adjustment should not need to be made in the position_nudge argument, e.g.
factor = 100
data <- data.frame(
name=c("de","gb","cn","ir","ru") ,
value=c(3,12,5,18,45)*factor
)
ggplot(data,
aes(x=name, y=value)) +
geom_bar(stat = "identity") +
geom_text(aes(y = 0,
label = paste0(name,value)),
position = position_nudge(y = -12)) +
coord_cartesian(ylim = c(0, 50*factor), # This focuses the x-axis on the range of interest
clip = 'off') + # This keeps the labels from disappearing
theme(plot.margin = unit(c(1,1,1,1), "lines"))
Currently, this does not work, so that I need to manually change -12 to -1200 to achieve this:
This is of course only a short reproducible example, the actual use-case is placing country flags as x-axis labels below the plot.
The final product will look somewhat like this, but currently requires updating the nudges each time the y-values change:
Thank you very much!

Answer 1

The easiest "hack" is to make this two plots and bind them with patchwork or cowplot. If you try it differently, you'd soon get into deep grid ... trouble.
Related
baptiste on github
baptiste on stackoverflow
Sandy Muspratt's answer
The easy way:
library(ggplot2)
library(patchwork)
foo <- data.frame(
name=c("de","gb","cn","ir","ru") ,
value=c(3,12,5,18,45)*1
)
foo_label = paste(foo$name, foo$value)
p <- ggplot(foo, aes(x=name, y=value)) +
geom_blank() # essential, so that both plots have same scaling
p_1 <-
p + geom_col() +
coord_cartesian(ylim = c(0, 50),clip = 'off') +
theme(plot.margin = margin())
p_text <-
p + annotate("text", label = foo_label, x = 1:5, y = 0, col="red") +
theme_void() +
coord_cartesian(clip = "off") +
theme(plot.margin = margin(1,0,1,0, unit = "lines"))
p_1/p_text + plot_layout(heights = c(1,0)) #this is a workaround to make the height of the text plot minimal!
You can then of course annotate with anything.

Answer 2

For your stated goal, the ggtext library may be more appropriate, as it allows you to embed images directly into the x axis labels. See also here for another example.
library(ggplot2)
library(ggtext)
labels <- c(
setosa = "<img src='https://upload.wikimedia.org/wikipedia/commons/thumb/8/86/Iris_setosa.JPG/180px-Iris_setosa.JPG'
width='100' /><br>*I. setosa*",
virginica = "<img src='https://upload.wikimedia.org/wikipedia/commons/thumb/3/38/Iris_virginica_-_NRCS.jpg/320px-Iris_virginica_-_NRCS.jpg'
width='100' /><br>*I. virginica*",
versicolor = "<img src='https://upload.wikimedia.org/wikipedia/commons/thumb/2/27/20140427Iris_versicolor1.jpg/320px-20140427Iris_versicolor1.jpg'
width='100' /><br>*I. versicolor*"
)
ggplot(iris, aes(Species, Sepal.Width)) +
geom_boxplot() +
scale_x_discrete(
name = NULL,
labels = labels
) +
theme(
axis.text.x = element_markdown(color = "black", size = 11)
)