Is there a way to plot an annotation_custom() at the same point of the plot despite the input dataset?

Question

I have a list of dataframes. The dataframes have the columns: Dist, GPSSpeed_kmh. I want to plot a png of the left-bend sign in a specific point of the plot. I used the following code for the dataframe df and the result is shown below:

  ggplot(df, aes(Dist, GPSSpeed_kmh)) +
    geom_point(size = 1) + 
    geom_smooth(se = TRUE, span = 0.7) +
    scale_x_continuous(limits = c(-20,11),breaks = seq(-30, 10, by = 10), labels = abs(breaks)) +
    geom_vline(xintercept = -16) +
    annotation_custom(leftbend, xmin = -18, xmax = -14, ymin = 30, ymax = 40)

But when I use another dataframe the leftbend icon is plotted in another point because the arguments ymin,ymax depend on the input dataframe in order to plot the leftbend icon above the geom_vline() (xmin, xmax are fixed). Is there a way for the arguments ymin,ymax to be adjusted in the input dataframe? I want to have a common output result for all my dataframes.

Thank you in advance!

*if I use another dataframe the result is the following:

The icon is even not plotted because this dataframe don't have GPSSpeed_kmh values from 30 (ymin) to 40 (ymax) as the other have!

Answer 1

What I have meant with my comment is the following:

library(ggplot2)
library(png)             ## for loading the png
library(grid)            ## to render the image
y=c(1,2,3)
x=c(0,0,0)
d=data.frame(x=x, y=y)   ## make sample dataframe

img <- readPNG(system.file("img", "Rlogo.png", package="png"))
g <- rasterGrob(img, interpolate=TRUE)

mi <- 1; ma <- 2.2       ## defining limits for y-axis
ggplot(d, aes(x, y)) + geom_point() + 
       ylim(mi, ma)+     ## set limits for y-axis
       annotation_custom(g, xmin=.020, xmax=.030, ymin=.5*ma) 

By relating ymin of annotation_custom to half the max (or whatever fraction you like) of the max of the y-axis, you make sure the image is always at the same position.

Further help can be found here and here.

I hope this helps.