#1

Sort of a followup to plot suggestions for two categorical vars while trying out geom_count.

Is there a way to make ggforce ignore factors when using geom_mark()? I'd like to highlight certain areas of the plot, not the individual points. Replacing factors with numeric kind of messes up the ggplot part of the plot.

Edit: @thomasp85, I hope you don't mind me tagging you here.

library(tidyverse)
library(ggforce)
#> 
#> Attaching package: 'ggforce'
#> The following objects are masked from 'package:ggplot2':
#> 
#>     stat_ellipse, StatEllipse

Df <- tibble::tribble(
  ~x, ~y,
  4,  2,
  1,  1,
  3,  2,
  1,  0,
  1,  0,
  3,  0,
  5,  2,
  5,  4,
  3,  2,
  3,  0,
  5,  2,
  4,  3,
  1,  0,
  4,  1,
  1,  0,
  1,  0,
  4,  1,
  3,  0,
  5,  3,
  1,  0,
  1,  0,
  1,  0,
  1,  0,
  4,  0,
  1,  0,
  1,  0,
  5,  5,
  1,  0,
  5,  0,
  3,  3,
  1,  2,
  4,  2,
  1,  0,
  5,  5,
  4,  4,
  5,  4,
  3,  0,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  5,  4,
  1,  0,
  1,  1,
  5,  3,
  5,  4,
  5,  5,
  1,  4,
  5,  3,
  1,  0,
  1,  0,
  5,  1,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  5,  2,
  1,  0,
  1,  0,
  1,  1,
  5,  3,
  1,  0,
  5,  4,
  1,  2,
  3,  1,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  1,  0,
  1,  0,
  5,  5,
  1,  1,
  1,  0,
  5,  3,
  1,  0,
  5,  2,
  1,  0,
  4,  1,
  5,  3,
  3,  0,
  3,  0,
  1,  0,
  5,  2,
  1,  3,
  5,  2,
  5,  5,
  5,  3,
  3,  2,
  1,  0,
  4,  2,
  5,  3,
  5,  4,
  3,  0,
  1,  0,
  5,  0
) %>% 
  mutate_if(is.numeric, ordered) %>% 
  mutate(cm = case_when(
    x >= 3 & y >= 2 ~ "TP",
    x == 1 & y <= 1 ~ "TN"
  ))
  
ggplot(Df, aes(x, y)) +
  geom_count(aes(color = y, size = stat(prop), group = y)) +
  scale_size_area(max_size = 25)



ggplot(Df, aes(x, y)) +
  geom_count(aes(color = y, size = stat(prop), group = y)) +
  scale_size_area(max_size = 25) +
  geom_mark_rect(aes(fill = cm, filter = cm == "TP"))



Df %>% 
  mutate(x = as.numeric(as.character(x)),
         y = as.numeric(as.character(y))) %>% 
  ggplot(aes(x, y)) +
    geom_count(aes(color = y, size = stat(prop), group = y)) +
    scale_size_area(max_size = 25) +
    geom_mark_rect(aes(fill = cm, filter = cm == "TP"))

Created on 2019-03-04 by the reprex package (v0.2.1)

#2

Simply add group = fill to the geom_mark aesthetics. When group is not set it is automatically created from the interaction of all non-numeric aesthetics (this is a ggplot2 convention — not ggforce), which is why you see different behaviour depending on whether x and y are factors or not

1 Like
#3

Thank you! That (almost) does it :slight_smile:

I'm hoping this is just as simple; how do I separate the groups when adding more than one group in filter? Note that I've added a few more groups in case_when.

library(tidyverse)
library(ggforce)
#> 
#> Attaching package: 'ggforce'
#> The following objects are masked from 'package:ggplot2':
#> 
#>     stat_ellipse, StatEllipse

Df <- tibble::tribble(
  ~x, ~y,
  4,  2,
  1,  1,
  3,  2,
  1,  0,
  1,  0,
  3,  0,
  5,  2,
  5,  4,
  3,  2,
  3,  0,
  5,  2,
  4,  3,
  1,  0,
  4,  1,
  1,  0,
  1,  0,
  4,  1,
  3,  0,
  5,  3,
  1,  0,
  1,  0,
  1,  0,
  1,  0,
  4,  0,
  1,  0,
  1,  0,
  5,  5,
  1,  0,
  5,  0,
  3,  3,
  1,  2,
  4,  2,
  1,  0,
  5,  5,
  4,  4,
  5,  4,
  3,  0,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  5,  4,
  1,  0,
  1,  1,
  5,  3,
  5,  4,
  5,  5,
  1,  4,
  5,  3,
  1,  0,
  1,  0,
  5,  1,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  5,  2,
  1,  0,
  1,  0,
  1,  1,
  5,  3,
  1,  0,
  5,  4,
  1,  2,
  3,  1,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  1,  0,
  1,  0,
  5,  5,
  1,  1,
  1,  0,
  5,  3,
  1,  0,
  5,  2,
  1,  0,
  4,  1,
  5,  3,
  3,  0,
  3,  0,
  1,  0,
  5,  2,
  1,  3,
  5,  2,
  5,  5,
  5,  3,
  3,  2,
  1,  0,
  4,  2,
  5,  3,
  5,  4,
  3,  0,
  1,  0,
  5,  0
) %>% 
  mutate_if(is.numeric, ordered) %>% 
  mutate(cm = case_when(
    x >= 3 & y >= 2 ~ "TP",
    x == 1 & y <= 1 ~ "TN",
    x >= 3 & y <= 1 ~ "FP",
    x == 1 & y >= 2 ~ "FN"
  ))


ggplot(Df, aes(x, y)) +
  geom_count(aes(color = y, size = stat(prop), group = y)) +
  scale_size_area(max_size = 25) +
  geom_mark_rect(aes(fill = cm, group = "fill", label = cm, filter = cm == "TP"))



ggplot(Df, aes(x, y)) +
  geom_count(aes(color = y, size = stat(prop), group = y)) +
  scale_size_area(max_size = 25) +
  geom_mark_rect(aes(fill = cm, group = "fill", label = cm, filter = cm == c("TP", "FN")))

Created on 2019-03-04 by the reprex package (v0.2.1)

#4

Nevermind...

geom_mark_rect(aes(group = cm, fill = cm))

I'll see myself out.

#5

My first message was a bit unclear — glad you figured it out

#6

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.