Help with marking on a ridge plot

[cmues]

Hi all,

I've been struggling with making a graph. In this ridgeplot, I would like to mark an X value, for each category on the Y axis.

I've been toying with geom_label. But this only gives me a label (well, its in the name I guess ). And this label somehow takes on one of the fill values.

So I suppose my question is; in the plot the code generates:

1. is it possible to add a marking on every ridge, corresponding to the "pop" value?
2. if not, can I let the label show up in white? Or in a different shape of box?

Thanks!

library(tidyverse)
library(ggridges)

dfs <-
  data.frame(
    "estimate" = rnorm(300),
    "loading" = factor(rep(1:5, 60)),
    "set" = factor(rep(1:3, 100),),
    "pop" = rep(c(0.1, 0.15, 0.20, 0.05, 0.7), 60)
  )

ggplot(dfs, aes(x = estimate, y = loading, fill = set)) +
  geom_density_ridges(
    jittered_points = TRUE,
    point_shape = "|", point_size = 2, size = 0.25,
    position = position_points_jitter(height = 0), alpha = 0.6, scale = 1.2) 
#> Picking joint bandwidth of 0.445

loadsummary2 <- crossing("set" =  dfs$set, "loading" = dfs$loading)
loadsummary2$pop <- rep(c(0.1, 0.15, 0.20, 0.05, 0.7), 3)
  
ggplot(dfs, aes(x = estimate, y = loading, fill = set)) +
  geom_density_ridges(
    jittered_points = TRUE,
    point_shape = "|", point_size = 2, size = 0.25,
    position = position_points_jitter(height = 0), alpha = 0.6, scale = 1.2) + 
    geom_label(data = loadsummary2, aes(y = loading, x = pop), label = loadsummary2$pop)
#> Picking joint bandwidth of 0.445

^{Created on 2020-03-13 by the reprex package (v0.3.0)}

[dromano]

Do you have an ideal location for where the labels should go? Like in the space above and to the left of the curves?

[cmues]

Ideally, it would look like this (the solution). I've played around with that code, and can't seem to adapt it.

I'm also just terrible with ggplot though, so I might be missing something simple.

[dromano]

So just to confirm I understand, you want labels at the the pop values themselves, then? (With those exact numbers in them, right?)

[cmues]

Yes, at those exact values!

[dromano]

Could you try using this to replace your current line with geom_label()?

geom_text(data = loadsummary2, 
          aes(y = loading, x = pop, label = pop),
          position=position_nudge(y=-0.1), 
          colour="red", 
          size=3.5)
)

I think you want the label argument inside of the aes() function.

[cmues]

Oh, something that silly!
Any solution for adding line segments?

[dromano]

Yes, you should be able to by using geom_segment(). Does this work?

geom_segment(
  aes(
    x = pop,
    y = loading,
    xend = pop,
    yend = loading + pop
  )
)

[dromano]

Sorry, the yend doesn't make sense...let me see if I can fix that.

[cmues]

@dromano

geom_segment( aes( x = pop, y = loading, xend = pop, yend = loading + pop ) )

Sadly, no.

Picking joint bandwidth of 0.448
Warning messages:
1: In Ops.factor(loading, pop) : ‘+’ not meaningful for factors
2: Removed 300 rows containing missing values (geom_segment).

I played around with making them not factors, but I still got the second error message.

[dromano]

I think you could try as.numeric(loading) + 1 (+ pop) doesn't make sense, and see if that at least makes vertical lines. Still thinking about the lines...

[dromano]

Sorry forgot to have you include the data in the geom_segment() layer. Here's the corrected version, but I have question for you: Did you want the lines to be a particular height?

  geom_segment( 
    data = loadsummary2, 
    aes( x = pop, y = loading, xend = pop, yend = as.numeric(loading) + 1 ) 
    )

[cmues]

I've been playing around. I can fudge with the numbers a little more, see what looks best. But for now this is more than sufficient to share as internal results!

Thanks so much for your effort!

p + geom_segment(aes(
  x = pop,, 
  y = as.numeric(loading) - .05,
  xend = pop,
  yend = as.numeric(loading) + .15))