rggplot2ggridges

ggridges - controlling legend

I have a ridge plot where I'm colouring points by a grouping that doesn't affect the ridges. I would like to have a continuous point_colour legend using viridis, but can't figure out how to control it. I have an example below showing how to control a discrete point_colour aesthetic and legend successfully. How do I do the same, but with a continuous point_colour aesthetic?

toy example:

iris %>% 
    mutate(Group = sample(1:4, n(), replace = TRUE)) %>%
    ggplot() +
    geom_density_ridges(aes(x = Sepal.Length, y = Species, point_colour = as.factor(Group), group = Species),
            rel_min_height = 0.005, jittered_points = TRUE) +
    scale_y_discrete(expand = c(0.01, 0)) +
    scale_x_continuous(expand = c(0.01, 0)) +
    theme_ridges() +
    scale_discrete_manual("point_colour",
                        values = viridis::viridis(n = 4,
                                                  begin = 0, end = 1, direction = 1),  name = "Group") +
    guides(point_colour = guide_legend(override.aes = list(fill = "white", colour = "white")))
Solution

  • The issue with the legend when using a continuous color scale for point_colour is that by default a GuideColourbar can only be used for the color or fill aes. But a possible fix would be to extend GuideColourbar by adding point_colour to the available aesthetics:

    library(ggplot2)
library(ggridges)
library(dplyr, warn = FALSE)

set.seed(123)

GuideColourbar2 <- ggproto(
  "GuideColourbar2", GuideColourbar,
  available_aes = c(
    GuideColourbar$available_aes,
    "point_colour"
  )
)

iris %>%
  mutate(Group = sample(1:4, n(), replace = TRUE)) %>%
  ggplot() +
  geom_density_ridges(
    aes(
      x = Sepal.Length, y = Species,
      point_colour = Group, group = Species
    ),
    rel_min_height = 0.005, jittered_points = TRUE
  ) +
  scale_y_discrete(expand = c(0.01, 0)) +
  scale_x_continuous(expand = c(0.01, 0)) +
  theme_ridges() +
  scale_color_viridis_c(
    aesthetics = "point_colour",
    guide = GuideColourbar2
  )
#> Picking joint bandwidth of 0.181