CODEWITHSUNDEEP
rdataframeggplot2
afleishman
afleishman

Reputation: 433

Create a lollipop chart to compare groups

I have the following data frame

library(ggplot2)

set.seed(149)

x <- data.frame(
  region = factor(rep(1:10, each = 2)),
  group = rep(c("O", "E"), 10),
  mean = sample(1:2, 20, replace = TRUE)
)

x

   region group mean
1       1     O    2
2       1     E    1
3       2     O    1
4       2     E    2
5       3     O    1
6       3     E    1
7       4     O    1
8       4     E    1
9       5     O    1
10      5     E    2
11      6     O    2
12      6     E    2
13      7     O    1
14      7     E    1
15      8     O    1
16      8     E    1
17      9     O    1
18      9     E    2
19     10     O    1
20     10     E    1

I'm trying to create a lollipop chart, so I can make simple comparisons between each region (either 'O' or 'E'). I'd make a dumbbell plot, but a lot of the region's two groups are identical so the dumbbell plot ends up looking a lot like a dot plot.

Here's what I have so far...

ggplot(x, aes(y = region, x = mean, label = mean, fill = group, colour = group)) +
  geom_segment(aes(x = 0, y = region, xend = mean, yend = region), color = "grey50", size = 0.75) +
  geom_point(size = 3) +
  geom_text(nudge_x = 1.5, angle = -45)

enter image description here

Basically, for each region I'd like to plot two lines, one for group 'O' and one for group 'E.'

Upvotes: 5

Views: 9981

Answers (2)

HAVB
HAVB

Reputation: 1888

Faceting can also accomplish the task, and may be particularly useful when you have a bunch of categories (i.e. not just two but 4, 6, etc)

ggplot(x, aes(y = region, x = mean, label = mean, fill = group, colour = group)) +
    geom_segment(aes(x = 0, y = region, xend = mean, yend = region), color = "grey50", size = 0.75) +
    geom_point(size = 3) +
    facet_wrap(~group)

enter image description here

Upvotes: 2

bouncyball
bouncyball

Reputation: 10771

We can accomplish this by using geom_linerange (the "stick"), geom_point (the "candy"), and by specifying position = position_dodge:

ggplot(x)+
    geom_linerange(aes(x = region, ymin = 0, ymax = mean, colour = group), 
                   position = position_dodge(width = 1))+
    geom_point(aes(x = region, y = mean, colour = group),
               position = position_dodge(width = 1))+
    coord_flip()

enter image description here

Upvotes: 12

Related Questions