stat = "density" and jittered_points do not work together

[akarlinsky]

From the vignette, it states that in order to compute the density separately for each group, we can use stat="density" and height=..density.. .

Unfortunately, It seems that this doesn't work with jittered points, as you can see here:

Jittered points work as expected, but the densities look too much alike as the density is estimated for all X:

library(tidyverse)
library(ggridges)

iris %>%
  mutate(group_pl = cut_number(Petal.Length, n=5)) %>%
  ggplot(aes(x = Sepal.Length, y = Species)) + 
  geom_density_ridges(jittered_points = TRUE, 
                      alpha = 0.8, 
                      scale = 1,
                      rel_min_height = 0.01,
                      aes(point_fill = group_pl, 
                          point_color = group_pl, 
                          group = Species)) +
  theme_ridges() +
  theme(legend.position = "none")
#> Picking joint bandwidth of 0.181

Using stat="density" and height = ..density.. gives the expected densities, but points are no longer available - parameter us "unknown".

iris %>%
  mutate(group_pl = cut_number(Petal.Length, n=5)) %>%
  ggplot(aes(x = Sepal.Length, y = Species)) + 
  geom_density_ridges(jittered_points = TRUE, 
                      alpha = 0.8, 
                      scale = 1,
                      rel_min_height = 0.01,
                      aes(point_fill = group_pl, 
                          point_color = group_pl, 
                          group = Species, 
                          height = ..density..),
                      stat = "density") +
  theme_ridges()
#> Warning: Ignoring unknown parameters: jittered_points

^{Created on 2019-09-17 by the reprex package (v0.3.0)}

[clauswilke]

See #42. stat_density() does not have the same parameter options that stat_density_ridges() has. I'm happy to accept a PR that improves the documentation.

[akarlinsky]

I thought adding an option to geom_density_ridges such as "bygroup" which flags if the bandwidth estimation should be by group or on the whole range of x values.

Unfortunately, I'm not skilled enough in R programming to achieve this.

[clauswilke]

Please file an issue with this specific request then. I'm going to close this one.