geom_text does not appear using scale_y_log10

Question

I can add text to a ggplot using geom_text just fine (example using the cars database):

ggplot(cars, aes(speed, dist), color='blue') + 
geom_point(size=0.8) + 
geom_text(y=25, x=5, aes(label=paste("sigma == 1"), size=1.5), 
    hjust=0, parse=TRUE, color='blue')

But when I change the y scale to logarithmic, I can not get the text to appear on the graph:

ggplot(cars, aes(speed, dist), color='blue') + 
geom_point(size=0.8) + 
geom_text(y=25, x=5, aes(label=paste("sigma == 1"), size=1.5), 
    hjust=0, parse=TRUE, color='blue') + 
scale_y_log10()

I've tried varying the text size and position but can't seem to get it.

Dekameter

Answer 1

This works for me:

require(ggplot2)
g <- ggplot(cars, aes(speed, dist), color='blue')
g <- g + geom_point(size=0.8)
g <- g + geom_text(aes(y=25, x=5, label=paste("sigma == 1"), size=1.5), hjust=0, parse=TRUE, color='blue')
g <- g + scale_y_log10()
g

Nymphet

Interesting and odd. That works for me. And now, when I reissue the commands as I did above, that too is working. Very strange. Thank you for your help. – Dekameter

Well then it was just the y=25 and x=5 being outside the aes which gave ggplot a hickup... – Nymphet

Answer 2

This works better:

ggplot(cars, aes(speed, dist), color='blue') + 
    geom_point(size=0.8) + 
    geom_text(y=2, x=5, aes(label=paste("sigma == 1"), size=1.5), 
        hjust=0, parse=TRUE, color='blue') + 
    scale_y_log10()

enter image description here

Normal to log scale: 100 -> 2; 1000 -> 3; 0.1 -> -1; 0.01 -> -2

Starveling

I like the idea of feeding the positions in on the scale of data and using the scales machinery to do the log transformations, rather than computing. – Instantaneous

Answer 3

The issue in the original code is that the x and y values are not wrapped up by aes(), so the axis transformation is not applied to them. To fix, just move the aes() to include all the aesthetics.

ggplot(cars, aes(speed, dist), color='blue') + 
  geom_point(size=0.8) + 
  geom_text(aes(y=25, x=5, label=paste("sigma == 1"), size=1.5), 
            hjust=0, parse=TRUE, color='blue') + 
  scale_y_log10()

Reckford

Even though the solution here is the same as the accepted answer, I think this is much better as it includes the explanation. – Instantaneous

Answer 4

Updating the solution with ggplot2 annotate() function to avoid warnings

library(ggplot2)
# Solution With annotate()
ggplot(cars, aes(speed, dist), color='blue')+
  geom_point(size=0.8)+
  annotate("text",y=25, x=5, label=paste("sigma == 1"), size=5, hjust=0, parse=TRUE, color='blue')+
  scale_y_log10()

The old way Note Warnings

# old way Note the annotation is included in legend 
#         Note the warning about Warning aesthetics

ggplot(cars, aes(speed, dist), color='blue')+
  geom_point(size=0.8)+
  geom_text(aes(y=25, x=5, label=paste("sigma == 1"), size=1.5), hjust=0, parse=TRUE, color='blue')+
  scale_y_log10()
#> Warning in geom_text(aes(y = 25, x = 5, label = paste("sigma == 1"), size = 1.5), : All aesthetics have length 1, but the data has 50 rows.
#> ℹ Did you mean to use `annotate()`?

^{Created on 2024-09-03 with reprex v2.1.0}

Young