CODEWITHSUNDEEP
rggplot2facet
banbh
banbh

Reputation: 1535

Adding multiple annotations when faceting

I'm having trouble adding multiple annotations (using vectors) to a plot with facets.

For example:

library(tidyverse) # ggplot2_3.3.0

tibble(t = 1:100) %>% 
  crossing(id = LETTERS[1:2]) %>% 
  group_by(id) %>% 
  mutate(y = cumsum(rnorm(n()))) %>% 
  ggplot(aes(t, y)) + # perhaps add `group = id` if you don't facet by `id`
  facet_wrap(vars(id)) + # (1)
  annotate('rect', xmin = 20, xmax = 30, ymin = -Inf, ymax = Inf, fill = 'grey60') + # (2)
  annotate('rect', xmin = 30, xmax = 40, ymin = -Inf, ymax = Inf, fill = 'grey70') + # (2)
  annotate('rect', xmin = 40, xmax = 50, ymin = -Inf, ymax = Inf, fill = 'grey80') + # (2)
  annotate('rect', xmin = 50, xmax = 60, ymin = -Inf, ymax = Inf, fill = 'grey90') + # (2)
  # annotate('rect', ymin = -Inf, ymax = Inf,                # (3)
  #          xmin = seq(20, by=10, len=4),                   # (3)
  #          xmax = seq(30, by=10, len=4),                   # (3)
  #          fill = paste0('grey', seq(60, by=10, len=4))) + # (3)
  geom_line() +
  theme_light()

The above code produces the desired plot (in particular, I want the same annotation on all facets). However, the annotate command is repeated four times; furthermore the help page for annotate says "the properties of the geoms are ... passed in as vectors". So a natural thing to try is to comment out lines (2), and uncomment lines (3). Unfortunately this generates the error

Error: Aesthetics must be either length 1 or the same as the data (8): fill

Note that if, in addition, you comment out line (1) (and optionally add group = id to the aesthetics) then it does not generate an error.

Upvotes: 2

Views: 512

Answers (1)

stefan
stefan

Reputation: 125797

For a discussion of this behaviour of ggplot2 see gihub. As the issue was closed and as far as I get it there is nothing you can do abozt that by using annotate. However, to achieve what you want you can simply use geom_rect like so:

library(tidyverse) # ggplot2_3.3.0

df_annotate <- data.frame(
  xmin = seq(20, 50, 10),
  xmax = seq(30, 60, 10),
  ymin = -Inf,
  ymax = Inf,
  fill = paste0("grey", seq(60, 90, 10))
)

tibble(t = 1:100) %>% 
  crossing(id = LETTERS[1:4]) %>% 
  group_by(id) %>% 
  mutate(y = cumsum(rnorm(n()))) %>% 
  ggplot() + 
  geom_rect(aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax, fill = fill), data = df_annotate) +
  geom_line(aes(t, y)) +
  scale_fill_identity() +
  facet_wrap(vars(id)) +
  theme_light()

Created on 2020-05-28 by the reprex package (v0.3.0)

Edit Using ggnewscale it's possible to have a second or ... fill scale:

library(tidyverse) # ggplot2_3.3.0
library(ggnewscale)

df_annotate <- data.frame(
  xmin = seq(20, 50, 10),
  xmax = seq(30, 60, 10),
  ymin = -Inf,
  ymax = Inf,
  fill = paste0("grey", seq(60, 90, 10))
)

df <- tibble(t = 1:100) %>% 
  crossing(id = LETTERS[1:4]) %>% 
  group_by(id) %>% 
  mutate(y = cumsum(rnorm(n())))

ggplot() + 
  geom_rect(aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax, fill = fill), data = df_annotate) +
  scale_fill_identity() +
  new_scale_fill() +
  geom_area(data = df, aes(t, y, fill = id)) +
  facet_wrap(vars(id)) +
  theme_light()

Created on 2020-05-29 by the reprex package (v0.3.0)

Upvotes: 2

Related Questions