Skip to content

Instantly share code, notes, and snippets.

What would you like to do?
Set continuous scales to be identical across ggplot2 plots automatically
# Check for a walkthrough.
# Note that this code is slightly "better" than the code in the post, since it checks
# to see if the scales being supplied and used are all continuous.
# Examples included after the function definitions.
# Given a bunch of plots and a scale you want to apply to them, this returns that scale,
# but with limits that encompass the union of all the ranges of values in those plots
# Meant to be for users
get_shared_scale <- function(..., scale) {
if (!inherits(scale$range, "RangeContinuous"))
stop("Supplied scale needs to be continuous")
plots <- list(...)
ranges <- purrr::map(plots, ~simple_range_extracter(., scale))
scale$limits <- range(unlist(ranges))
# Given a plot and a scale, get the range of the values used in the scale for that plot
# Not really intended for users
simple_range_extracter <- function(p, scale) {
d <- ggplot2::ggplot_build(p)
p_range <- d$plot$scales$get_scales(scale$aesthetics)$range
if (!inherits(p_range, "RangeContinuous"))
stop("Plot's scale for `", scale$aesthetics,
"` is not continuous")
# Given the unquoted variable names of a bunch of plots that have been
# saved to the current environment, and a scale that you want to apply
# to them, this will call `get_shared_scale()` using those plots and scale
# add that scale to the plots, and then assign these new plots to those
# variable names, essentially editing the plots 'in place'
set_scale_union <- function(..., scale) {
exprs <- rlang::enexprs(...)
scale <- get_shared_scale(..., scale = scale)
var_nms <- purrr::map_chr(exprs, rlang::as_name)
edit_plots_in_place(var_nms, env = parent.frame(),
scale = scale)
# Invisibly return the scale, in case you need it later
# A sub-function. *Definitely* not intended for users
edit_plots_in_place <- function(names, env, scale) {
vars <- rlang::env_has(env = env, nms = names)
if (!all(vars))
stop("Environment does not have variables for ",
paste(names(vars[!vars]), collapse=", "))
purrr:::walk(names, function(nm) {
og_plot <- rlang::env_get(env, nm = nm)
message("Changing plot `", nm, "`")
# Muffles messages about already having scales
assign(x = nm, envir = env,
value = og_plot + scale),
message = function(err) {
if (grepl("already present", err$message))
# Examples -------------------------------------------------
# Make the plots
p1 <- data.frame(x = runif(400),
y = runif(400)) %>%
ggplot(aes(x, y)) +
stat_density_2d(geom = "raster",
aes(fill = after_stat(density)),
contour = FALSE) +
scale_fill_viridis_c("z") +
theme(legend.position = "bottom")
p2 <- data.frame(x = c(runif(200), runif(200, 0.25, 0.75)),
y = c(runif(200), runif(200, 0.25, 0.75))) %>%
ggplot(aes(x, y)) +
stat_density_2d(geom = "raster",
aes(fill = after_stat(density)),
contour = FALSE) +
scale_fill_viridis_c("z") +
theme(legend.position = "bottom")
# The color scales don't match:
# I'm setting the scale I want to eventually use as a variable
# so I don't have to keep retyping it for these examples
my_viridis_scale <- scale_fill_viridis_c(name = "z")
# Test the range:
simple_range_extracter(p1, my_viridis_scale)
# Get the shared scale
shared_scale <- get_shared_scale(p1, p2, scale=my_viridis_scale)
# Test it out
p1 + shared_scale
p2 + shared_scale
# Now edit the plots in place
set_scale_union(p1, p2, scale=my_viridis_scale)
# Now they match:
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.