Change y-axis tick mark labels manually with factor variables R

Question

I have the following data:

id <- rep(1:100)
A <- rep(c(0.12 ,0.25, 0.5, 1, 2), each = 20)
B <- rep(c(0.06, 0.03, 0.015, 0.12), each = 25)
C <- rep(c(0.015, 0.03, 0.06, 0.12, 0.25), each = 20)

df <- data.frame(id,A,B,C,stringsAsFactors = F)

I gather A,B, and C into two columns. Note that column A, B and C are actually factors, I just avoid specifying them as factors to create the violin plot.

library(dplyr)

df_edited <- df %>%
  gather(key, value, -id, factor_key = F)

I create the following plot with this data:

library(ggplot2)

factor_breaks <- c(0.015,0.03,0.06,0.12,0.25,0.5,1,2)
factor_levels <- c("0.015","0.03","0.06","0.12","0.25","0.5","1","2")

ggplot(df_edited, aes(key, value))+
  geom_violin()+
  scale_y_continuous(labels = factor_levels, breaks = factor_breaks)

This creates the following plot:

Is it possible to make the y-axis labels to be safely placed evenly, as in the plot below, and still make sure that the violin-plot is correct?

ggplot(df_edited, aes(key, factor(value)))+
  geom_violin()

Answer 1

While log10 would also work in this example, I think it is more generic to "rescale" the values using sapply:

df_edited$newValue <- sapply(as.character(df_edited$value), 
                             function(x){which(factor_levels == x)})

This basically returns the position of the values within the factors_levels vector.

ggplot(df_edited, aes(key, newValue))+
  geom_violin() + 
  scale_y_continuous(breaks = 1:length(factor_levels), 
                     labels = factor_levels)

Answer 2

One option is to transform the values of the y-axis

ggplot(df_edited, aes(key, value))+
 geom_violin()+
 scale_y_continuous(labels = factor_levels,
                    breaks = factor_breaks,
                    trans = 'log10')