CODEWITHSUNDEEP
rggplot2
FlxZer
FlxZer

Reputation: 23

R ggplot dual X axis, one linear and one log, but for different variable

I'd like to plot the sample dataset shown below like so: Y axis - Var2 (linear scale); X_bottom_axis - Var1 (linear scale); X_top_axis - Var3 (log scale)

The problem is that all my top X-axis is cramming together all the ticks and doesn't function as log scale. The picture exemplifies the format I want to achieve for my plot. Any ideas? Desirable format

Below code shows a sample dataset and my unsuccessful attempt to plot it

d1 <- data.frame(var1=c(1,2,3,4,5),var2=c(100,110,99,94,100),var3=c(1e-6,12e-6,27e-6,1e-5,1e-3))
ggplot(d1,aes(y=var2))+geom_point(aes(x=var1),col="blue")+geom_point(aes(x=var3),col="red")+scale_x_continuous(sec.axis = sec_axis(~., breaks = c(1e-6,1e-5,1e-4,1e-3,1e-2,1e-1,0)))

Upvotes: 0

Views: 94

Answers (1)

Allan Cameron
Allan Cameron

Reputation: 174348

Adding a secondary axis doesn't change where the red dots are plotted. You need to apply a transformation to the data so that they are approximately on the same scale as the blue dots. For example. taking log10 of var3 and adding 6 will change the red dots to be between the values of 1 and 3 so they are clearly visible on the plot. All the secondary axis does is to provide an annotation showing how these red dots should be interpreted. To do this, you need to pass a function that inverts this transformation (i.e. ~10^(.x - 6)

ggplot(d1, aes(y = var2)) +
  geom_point(aes(x = var1), col = "blue") +
  geom_point(aes(x = log10(var3) + 6), col = "red")+
  scale_x_continuous(sec.axis = sec_axis(~10^(.x - 6), breaks = 10^(-6:0))

enter image description here

Upvotes: 2

Related Questions