In much the same way step_log() has a signed argument, I think step_sqrt() should as well. A signed sqrt() would uphold the ordering of near-zero values better then what signed log would do
values <- c(-3 ^ seq(5, -3, by = -0.1), 0, 3 ^ seq(-3, 5, by = 0.1))
plot(values)
In much the same way
step_log()
has a signed argument, I thinkstep_sqrt()
should as well. A signedsqrt()
would uphold the ordering of near-zero values better then what signed log would doCreated on 2022-03-26 by the reprex package (v2.0.1)