ggplot and scale_x_continuous

Question

here a sample of data :

library(ggplot2)
lbreaks=c(-800,-600,-400,-300,-200,-100,-75,-50,-25,0,25,50)
ddf=data.frame(x=lbreaks,
y=seq(1,length(lbreaks),1))

and the scatter plot corresponding

I would use a custom scale (log or something similar for positive and negative values) for x-axis values to generate the same scatterplot with the equally spaced interval of x values. I tried to use several function in scale_x_continuous(trans = without sucess. How could I do this ? Many thanks for your help,

Answer 1

If you want the x axis breaks to be evenly spaced, the simplest thing to do is:

ggplot(ddf, aes(factor(x), y)) + 
  geom_point(colour = 'red') +
  labs(x = 'x')

Perhaps less flippantly, your specific transformation can only really be achieved accurately by a piecewise transformation function:

my_trans <- scales::trans_new('custom',
                      transform = function(x) {
                        dplyr::case_when(
                          x < -400 ~ (x + 800) / 200 + 1,
                          x < -100 ~ (x + 400) / 100 + 3,
                          TRUE ~ (x + 100) / 25 + 6)
                      },
                      inverse = function(x) {
                        dplyr::case_when(
                          x < 3 ~ (x - 1) * 200 - 800,
                          x < 6 ~ (x - 3) * 100 - 400,
                          TRUE ~ (x - 6) * 25 - 100)
                      })


ggplot(ddf, aes(x, y)) + 
  geom_point(colour = 'red') +
  scale_x_continuous(trans = my_trans, breaks = lbreaks)

This allows arbitrary data to be passed:

ddf=data.frame( x=rnorm(100,mean=-500,sd=200), y=rnorm(100,mean=0.5,sd=0.2) )

ggplot(ddf, aes(x, y)) + 
  geom_point(colour = 'red') +
  scale_x_continuous(trans = my_trans, breaks = lbreaks)

However, such arbitrary x axis manipulations are not ideal, as they can be misleading.