Search code examples
rggplot2scalefacet-wrapyaxis

Free y axis in only some subplots with facet_wrap

I am trying to plot three different subplots within the same graphic using facet_wrap. I know the y axis can be settled "free" or "fixed" with the argument scales of facet_wrap(). However, due to the characteristics of my data, I would like to keep the y axis fixed for the first two plots but re-scaled for the third one. I could do three individual plots and arrange them together, but as I am using this code inside a Shiny app this is not the most convenient. Below you will find a minimal example of my data and my code so far.

    ##Data example:
    Color Time  Metric
1   Green    0  200.00
2   Green    2  300.00
3   Green    4  600.00
4   Green    6  800.00
5   Green    8 1400.00
6   Green   10 2600.00
7     Red    0  150.00
8     Red    2  260.00
9     Red    4  400.00
10    Red    6  450.00
11    Red    8  600.00
12    Red   10  650.00
13 Yellow    0    0.10
14 Yellow    2    0.20
15 Yellow    4    0.30
16 Yellow    6    0.60
17 Yellow    8    0.55
18 Yellow   10    0.70
    
#With free scales
ggplot(Example) + geom_point(aes(x = Time, y = Metric)) + facet_wrap(facets = "Color", scales = "free")

#With fixed scales:
ggplot(Example) + geom_point(aes(x = Time, y = Metric)) + facet_wrap(facets = "Color", scales = "fixed")

As you see, neither alternative is really useful: if I set the scales as free I lose the comparison between Green and Red, which are approximately within the same ranges. However, if I set them as fix then it is impossible to see what is going on with Yellow. The ideal would be to have Green and Red on the same scale, and yellow in its own scale. Can this be done with facet_wrap() ?

Solution

  • Consider this like an option for you, but it uses a variant of facet_grid() found in package facetscales which allows you to define specific scales. Here the code:

    #Install
#devtools::install_github("zeehio/facetscales")
library(tidyverse)
library(facetscales)
#Define scales
scales_y <- list(
  `Green` = scale_y_continuous(limits = c(0, 2700)),
  `Red` = scale_y_continuous(limits = c(0, 2700)),
  `Yellow` = scale_y_continuous(limits = c(0, 0.7))
)
#Code
ggplot(Example) +
  geom_point(aes(x = Time, y = Metric)) +
  facet_grid_sc(Color~.,
                scales = list(y = scales_y))

    Output:

    enter image description here

    Some data used:

    #Data
Example <- structure(list(Color = c("Green", "Green", "Green", "Green", 
"Green", "Green", "Red", "Red", "Red", "Red", "Red", "Red", "Yellow", 
"Yellow", "Yellow", "Yellow", "Yellow", "Yellow"), Time = c(0L, 
2L, 4L, 6L, 8L, 10L, 0L, 2L, 4L, 6L, 8L, 10L, 0L, 2L, 4L, 6L, 
8L, 10L), Metric = c(200, 300, 600, 800, 1400, 2600, 150, 260, 
400, 450, 600, 650, 0.1, 0.2, 0.3, 0.6, 0.55, 0.7)), class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", 
"14", "15", "16", "17", "18"))