Same y-axis scale with ggarrange - r

Question

I would like to use ggarrange to combine two plots into a single one, and use the same scaling for the y-axis to make it easier to compare the values, in my case, temperature calculations.
This is my sample data:
Sample Group Temperature_A Temperature_B
1 Sample01 A 20 34
2 Sample02 B 21 31
3 Sample03 A 25 35
4 Sample04 A 18 27
5 Sample05 B 19 29
6 Sample06 A 22 32
7 Sample07 B 23 33
8 Sample08 B 17 28
9 Sample09 A 15 26
10 Sample10 B 28 39
Using the following code, I get the following plot:
p1 <- ggplot(df, aes(Group, Temperature_A)) +
geom_boxplot()
p2 <- ggplot(df, aes(Group, Temperature_B)) +
geom_boxplot()
plate <- ggarrange(p1, p2, align = "hv")
plate
What I would like to have is for them to share their y-axis so that I can easily compare the two temperature calculation methods. I can do this manually with + scale_y_continuous(limits = c(a,b)) where I set the appropriate values for a and b, however, I have a lot of different dataframes with different temperature ranges. Ideally, I would use the lowest value (+ some space) from both plots, and the highest value (+ some space) from both plots for the limits of both plots. Is there a way to achieve this?
My manual approach and the desired output:
p1 <- ggplot(df, aes(Group, Temperature_A)) +
geom_boxplot() +
scale_y_continuous(limits = c(10, 40))
p2 <- ggplot(df, aes(Group, Temperature_B)) +
geom_boxplot() +
scale_y_continuous(limits = c(10, 40))
plate <- ggarrange(p1, p2, align = "hv")
plate

Answer 1

This is just to illustrate how you could use facets. Because you have two columns for temperature, this is called the 'wide format'. If you reshape the data to the long format, you can easily use facets.
library(ggplot2)
df <- read.table(text = " Sample Group Temperature_A Temperature_B
1 Sample01 A 20 34
2 Sample02 B 21 31
3 Sample03 A 25 35
4 Sample04 A 18 27
5 Sample05 B 19 29
6 Sample06 A 22 32
7 Sample07 B 23 33
8 Sample08 B 17 28
9 Sample09 A 15 26
10 Sample10 B 28 39", header = TRUE)
# Reshape
long <- tidyr::pivot_longer(
df, c("Temperature_A", "Temperature_B"),
values_to = "Temperature", names_to = "Temp_name"
)
# Example of facets
ggplot(long) +
geom_boxplot(aes(Group, Temperature)) +
facet_wrap(~ Temp_name)
Created on 2021-07-27 by the reprex package (v1.0.0)

Answer 2

I agree with teunbrand. Here is an slightly modified approach:
bring data in long format
ignore outlieres outlier.shape = NA
adapt ylim and
facet_wrap
library(tidyverse)
df1 <- df %>%
pivot_longer(
cols = starts_with("Temperature"),
names_to = "Temperature",
values_to = "values"
)
ggplot(df1, aes(Group, values)) +
geom_boxplot(outlier.shape = NA) +
coord_cartesian(ylim = c(10, 40)) +
facet_wrap(~Temperature) +
theme_bw()

Answer 3

Specifically for J.Doe who appeared unsatisfied with the current accepted answer, I'll post another answer which again ignores the 'ggarrange' part, but generalises better. Instead, the {patchwork} package is used.
Notably, you can use & scale_y_continuous(...) to apply a scale to all previous plots. You can use the plot + plot syntax to combine plots.
library(ggplot2)
library(patchwork)
df <- read.table(text = " Sample Group Temperature_A Temperature_B
1 Sample01 A 20 34
2 Sample02 B 21 31
3 Sample03 A 25 35
4 Sample04 A 18 27
5 Sample05 B 19 29
6 Sample06 A 22 32
7 Sample07 B 23 33
8 Sample08 B 17 28
9 Sample09 A 15 26
10 Sample10 B 28 39", header = TRUE)
p1 <- ggplot(df, aes(Group, Temperature_A)) +
geom_boxplot()
p2 <- ggplot(df, aes(Group, Temperature_B)) +
geom_boxplot()
p1 + p2 & scale_y_continuous(limits = c(10, 40))
The same works if you have a list of plots and use wrap_plot().
plot_list <- list(p1, p2)
wrap_plots(plot_list) & scale_y_continuous(limits = c(10, 40))
Created on 2022-12-08 by the reprex package (v2.0.0)