Is there a possibility to add a third y-axis to a plot with ggplot2? I have three different datasources I want to display in the plot. I already added a second y-axis, for the next dataset the scale is again very different, why I'm looking now for a solution.
So far I only found how to add a second axis, for example as shown [here].(https://r-graph-gallery.com/line-chart-dual-Y-axis-ggplot2.html) But I whether there is an possibility to add another one... Thank you!
This is equally clunky, but shows how it can be done from scratch using only CRAN resources.
library(cowplot)
library(patchwork)
p1 <- ggplot(df, aes(Sepal.Width, Sepal.Length)) +
geom_point() + theme(axis.line = element_line())
p2 <- ggplot(df, aes(Sepal.Width, Petal.Width)) + geom_point() +
theme(axis.line = element_line())
p3 <- ggplot(df, aes(Sepal.Width, Petal.Length)) +
geom_point(aes(color = "Petal.Length")) +
geom_point(aes(y = Sepal.Length/100, color = "Sepal.Length")) +
geom_point(aes(y = Petal.Width / 1000, color = "Petal.Width")) +
theme(axis.line = element_line(),
plot.margin = margin(10, 10, 10, 30))
wrap_elements(get_plot_component(p1, "ylab-l")) +
wrap_elements(get_y_axis(p1)) +
wrap_elements(get_plot_component(p2, "ylab-l")) +
wrap_elements(get_y_axis(p2)) +
p3 +
plot_layout(widths = c(3, 1, 3, 1, 40))
Data used
df <- iris
df$Sepal.Length <- df$Sepal.Length * 100
df$Petal.Width <- df$Petal.Width * 1000
dput(df) (where df is the name of your own data frame.
This is a very clunky solution based on extracting elements from previously plotted graphs and editing grid objects. It may or not give you a workable solution.
source("https://raw.githubusercontent.com/davidearn/plague_growth/master/analysis/plots/3axes.R")
set.seed(101)
dd <- data.frame(x=rnorm(20),y=rnorm(20))
library(ggplot2)
gg0 <- ggplot(dd)
g1A <- gg0 + geom_point(aes(x,y))
g1B <- gg0 + geom_point(aes(x,10*y))
g1C <- gg0 + geom_point(aes(x,100*y))
## use return_gtable = TRUE if planning to add further axes
g2 <- combine_axes(g1A,g1B,add_pos="l", return_gtable=TRUE)
g3 <- combine_axes(g2,g1C,add_pos="l")
print(g3)
combine_axes function does not work for facet plots and adding axes to the bottom?
There is a solution that doesn’t make another axis but manipulates the labels on the second axis as HTML using the ggtext package.
First you find the coefficients to rebase the other 2 variables.
## Packages
library(ggtext);library(tidyverse)
## Data
df <- data.frame(category = paste0(rep("type_", 30), c("A", "B", "C")),
x_val = rep(1:10, 3), y_val = abs(rnorm(30))*c(100,10,1))
## Rebase
rebase_coef <- df %>% group_by(category) %>% summarise(rebase_coef = max(y_val)) %>%
ungroup() %>% mutate(rebase_coef = rebase_coef / max(rebase_coef))
df_rebase <- df %>%
mutate(y_val = case_when(category == "type_B" ~ y_val/rebase_coef$rebase_coef[2],
category == "type_C" ~ y_val/rebase_coef$rebase_coef[3],
TRUE ~ y_val))
Then, create a data frame that will make your labels from a vector that will e used as the primary axis breaks (I use a function that easily converts labels into HTML):
## Y Axis Breaks
vbreaks <- max(df$y_val)*c(1, 0.75, 0.5, 0.25, 0)
## Category colors
vcolor_cat <- c("#0077B7","#8B4500","#4DAF4A"); names(vcolor_cat) <- unique(df$category)
## HTML function
html_fun <- function(x, color = 'white', size = 12) {
paste0("<span style = 'color:", color,";font-size:", size, "pt'><b>", x, "</b></span>")
}
## 2nd Labels
axis_labs <- data.frame(y_A = vbreaks, y_B = vbreaks/10, y_C = vbreaks/100) %>%
mutate(ylab_txt = paste0(html_fun(round(y_B,2), color = vcolor_cat[2]), "<br>",
html_fun(round(y_C,2), color = vcolor_cat[3]) ))
Then you plot:
df_rebase %>% ggplot(aes(x = x_val, y = y_val, color = category)) + geom_point() +
scale_color_manual(values = vcolor_cat) +
scale_y_continuous(breaks = vbreaks, labels = function(x) round(x, 2), name = "1 Thing",
sec.axis = sec_axis(~.*1, labels = axis_labs$ylab_txt, breaks = vbreaks, name = "2 Things")) +
theme(
axis.title.y.left = element_textbox_simple(
hjust = 0.5, vjust = 0.5, halign = 0.5, valign = 0.5,
orientation = "left-rotated", width = 0.75, fill = "grey90",
padding = margin(4, 2, 3, 2), margin = margin(2, 2, 2, 0)),
axis.title.y.right = element_textbox_simple(
hjust = 0.5, vjust = 0.5, halign = 0.5, valign = 0.5,
orientation = "right-rotated", width = 0.75, fill = "grey90",
padding = margin(4, 2, 3, 2), margin = margin(2, 0, 2, 2)),
axis.text.y.left = element_markdown(
colour = vcol_metric[1], fill = 'transparent', size = 12, face = "bold",
hjust = 1, vjust = 0.5, halign = 1, valign = 0.5, lineheight = 0.75,
padding = ggplot2::margin(3, 1, 1, 1), margin = ggplot2::margin(0, 0, 0, 0),
align_widths = FALSE, align_heights = FALSE
),
axis.text.y.right = element_markdown(
fill = 'transparent', hjust = 0, vjust = 0.5, halign = 0, valign = 0.5, lineheight = 0.75,
padding = ggplot2::margin(3, 1, 1, 1), margin = ggplot2::margin(0, 0, 0, 0),
align_widths = FALSE, align_heights = FALSE
))
dput(data)? So we can help you better