Set the order of a stacked bar chart by the value of one of the variables

Question

I've been asked to produce a stacked bar chart with bars and values that are stacked and ordered in a precise manner.

In this case "A3" on the left, "A2" in the middle and "A1" on the right. I've got that covered.

The bit that escapes me is that I've also been asked to order the bars descending order by the value of "A1". In this case that would mean 'Value 11' appearing at the top, descending in order to 'Value 6'. The ordering of the bars is recorded in the vector 'plotOrder', which I've tried to use to achieve something in my current attempt below.

In trawling the many questions posed on SO about ordering stacked bar charts I have not found a question or an answer that describes or solves this particular problem. Any assistance would be very much appreciated.

The simplified code below downloads some random data that is in the same form with which I'm currently working with and produces the chart below.

library(ggplot2)

stackedBarPlot <- 
        ggplot(data) +
        aes(x = ValueName, y = Percent, fill = reorder(Response, plotOrder)) +
        geom_bar(position = "fill", stat = "identity") +
        coord_flip()

Which produces this chart:

Data:

structure(list(ValueName = c("Value 11", "Value 14", "Value 13", 
"Value 12", "Value 10", "Value 1", "Value 5", "Value 9", "Value 4", 
"Value 7", "Value 8", "Value 3", "Value 15", "Value 2", "Value 6", 
"Value 11", "Value 14", "Value 13", "Value 12", "Value 10", "Value 1", 
"Value 5", "Value 9", "Value 4", "Value 7", "Value 8", "Value 3", 
"Value 15", "Value 2", "Value 6", "Value 11", "Value 14", "Value 13", 
"Value 12", "Value 10", "Value 1", "Value 5", "Value 9", "Value 4", 
"Value 7", "Value 8", "Value 3", "Value 15", "Value 2", "Value 6"
), plotOrder = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 
12L, 13L, 14L, 15L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 
11L, 12L, 13L, 14L, 15L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 
10L, 11L, 12L, 13L, 14L, 15L), Response = c("A1", "A1", "A1", 
"A1", "A1", "A1", "A1", "A1", "A1", "A1", "A1", "A1", "A1", "A1", 
"A1", "A2", "A2", "A2", "A2", "A2", "A2", "A2", "A2", "A2", "A2", 
"A2", "A2", "A2", "A2", "A2", "A3", "A3", "A3", "A3", "A3", "A3", 
"A3", "A3", "A3", "A3", "A3", "A3", "A3", "A3", "A3"), Percent = c(86.5, 
85.3, 84.5, 83.2, 81, 79.5, 77, 76.7, 71, 66.2, 64.5, 60.5, 59.6, 
57.2, 53.2, 9.9, 9.4, 10.2, 9.9, 11.8, 14.7, 13.9, 13.5, 15.1, 
16.1, 21.3, 21.3, 26.6, 19.8, 18.5, 3.6, 5.3, 5.3, 6.9, 7.2, 
5.8, 9, 9.8, 13.9, 17.7, 14.1, 18.2, 13.8, 22.9, 28.2)), .Names = c("ValueName", 
"plotOrder", "Response", "Percent"), row.names = c(NA, -45L), class = c("tbl_df", 
"tbl", "data.frame"))

Thank you in advance for any advice or solutions you might have.

It is rare to see somebody pay so much attention to the way the code is written (IMO very pretty). Can you also make it self-contained (i.e. no links to off-site resources)? — Roman Luštrik, Jun 22 '17 at 07:42
I've added the data and simplified the code to reflect, hope that's ok — Axeman, Jun 22 '17 at 07:47

score 7 · Accepted Answer · answered Jun 22 '17 at 08:18

7

You already have plotOrder, but you need to apply it to reorder x rather than fill...

stackedBarPlot <- 
  ggplot(data) +
  aes(x = reorder(ValueName,-plotOrder), 
      y = Percent, 
      fill = Response) +
  geom_bar(position = "fill", stat = "identity") +
  coord_flip()

answered Jun 22 '17 at 08:18

Andrew Gustar

17,295
1
22
32

1

Thanks @Andrew Gustar that's an elegant solution. Works a treat. Very much appreciated! – vengefulsealion Jun 23 '17 at 02:30
@Andrew Gustar Great solution! Do you have any suggestions how to modify your code to achieve the same result, but lacking the plotOrder column? – M.Teich Sep 04 '18 at 13:35
4

@M.Teich It is a bit messy, but you could change the `x` aesthetic to something like `x = factor(ValueName,levels=ValueName[Response=="A1"][order(Percent[Response=="A1"])])` – Andrew Gustar Sep 10 '18 at 16:11

score 5 · Answer 2 · answered Jun 22 '17 at 07:58

With a little pre-calculation of the order, this is not so hard:

library(magrittr)
library(dplyr)

o <- d %>% filter(Response == "A1") %>% arrange(Percent) %>% extract2("ValueName")

d %>% 
  mutate(ValueName = factor(ValueName, o)) %>% 
  ggplot() +
  aes(x = ValueName, y = Percent, fill = reorder(Response, plotOrder)) +
  geom_bar(position = "fill", stat = "identity") +
  coord_flip()

Set the order of a stacked bar chart by the value of one of the variables

2 Answers2

Linked

Related