oob=scales::squish 不适用于 ggplot2 中的 scale_size_continuous

Question

This is in relation to this question here: Adjust geom_point size so large values are plotted, but do not appear larger in ggplot2?

Specifically in reference to the response by @Axeman

I couldn't comment on that question, and so had to ask a new question.

I wish to achieve the "squish"ing for the point size of geom_point , but the option oob=scales::squish doesn't work with scale_size_continuous . I am not sure what else am I missing.

Would appreciate any help. Here is the code I tried:

xx = ggplot(pcm, aes(x = variable, y = TF)) + 
  geom_point(aes(size = value, fill=value), shape = 21) + 
  scale_size_continuous(range=c(1, 12),
    limits = c(-2, 2),
    oob = scales::squish)

Further, I want to add that I cannot use scale_size_area as answered by @Axeman because I do not want 0 values to be mapped to points with size 0. The range of my data is approx. from -1.7 to +3. I want the smallest size allocated for the lowest negative value. Thanks.

Answer 1

Is it a problem if you handle it with a transformation on value ?

Here I just imagined how your data may look like:

pcm <- data.frame(variable = runif(100),
                  TF       = runif(100),
                  value    = runif(100, -1.7, 3))

Here is the plot

library(ggplot2)

ggplot(pcm, aes(x = variable, y = TF)) + 
    geom_point(aes(size = pmax(pmin(value, 2), -2), 
                   fill = value), 
               shape = 21) +
    labs(size = "value")

Let me know, I may be able to improve my answer.

Also I would suggest you just a couple of updates to improve the readability of your chart, but it's up to you.

ggplot(pcm, aes(x = variable, y = TF)) + 
    geom_point(aes(size = pmax(pmin(value, 2), -2), 
                   colour = value), 
               alpha = 0.6) +
    labs(title = "Dots and colours",
         size = "value") +
    theme_minimal()