Add inverse hyperbolic sine transformation

[frederikziebell]

Sometimes it's nice to have a 'log'-transformation that also works for negative values to have points more evenly spread-out:

ihs <- function(x){log(x+sqrt(x^2+1))}
ihs_trans <- scales::trans_new(name="ihs",transform=ihs, inverse=sinh)

set.seed(1)
data <- tibble(
  x=runif(10^4),
  y=rnorm(10^4)
)

data %>% 
  ggplot(aes(x,y)) +
    geom_point(alpha=.7) +
    ggtitle("regular")

data %>% 
  ggplot(aes(x,y)) +
    geom_point(alpha=.7) +
    scale_y_continuous(trans=ihs_trans) +
    ggtitle("ihs")

[hadley]

Sure, do you want to do a PR?

[thomasp85]

Fixed by #360