16

does anyone know if it is possible to exclude zero values from a barplot in ggplot?

I have a dataset that contains proportions as follows:

 X5employf       prop X5employff
1   increase 0.02272727           
2   increase 0.59090909          1
3   increase 0.02272727   1  and 8
4   increase 0.02272727          2
5   increase 0.34090909          3
6   increase 0.00000000          4
7   increase 0.00000000          5
8   increase 0.00000000          6
9   increase 0.00000000    6 and 7
10  increase 0.00000000   6 and 7 
11  increase 0.00000000          7
12  increase 0.00000000          8
13  decrease 0.00000000           
14  decrease 0.00000000          1
15  decrease 0.00000000   1  and 8
16  decrease 0.00000000          2
17  decrease 0.00000000          3
18  decrease 0.10000000          4
19  decrease 0.50000000          5
20  decrease 0.20000000          6
21  decrease 0.00000000    6 and 7
22  decrease 0.00000000   6 and 7 
23  decrease 0.10000000          7
24  decrease 0.10000000          8
25      same 0.00000000           
26      same 0.00000000          1
27      same 0.00000000   1  and 8
28      same 0.00000000          2
29      same 0.00000000          3
30      same 0.21052632          4
31      same 0.31578947          5
32      same 0.26315789          6
33      same 0.15789474    6 and 7
34      same 0.00000000   6 and 7 
35      same 0.05263158          7
36      same 0.00000000          8

as you can see in the 'prop' column there are a lot of zero values. I am producing a facetted bar plot with 'X5employf' column as the facet. But because of the zero values I end up with a lot of empty space on my plot(see below). Is there a way of forcing ggplot to not plot the zero values? Its not the case of dropping unused factors as these are not NA values but 0s. Any ideas??!

example plot

Improve this question

2 Answers 2

Reset to default
15

For your plot, simply use which to specify that you only want to use the subset of the dataframe containing non-zero proportions. This way you don't have to modify your original dataframe. Then, specify "free_x" in your scales argument within facet_grid to get rid of your empty space in your faceted plot.

plot <- ggplot(df[which(df$prop>0),], aes(X5employff, prop)) +
  geom_bar(aes(fill=X5employff, stat="identity")) +
  facet_grid( ~ X5employf, scales="free_x") +
  theme_bw()
plot

enter image description here

Note that I replaced the blank fields with "blank" for the sake of quick import into R from Excel.

Improve this answer
Sign up to request clarification or add additional context in comments.

4 Comments

Also, the answer to this similar question shows a slightly different way to do this, by adding drop=TRUE within the aes argument inside ggplot()
thanks sc_evans, that's what I wanted. still trying to get my head around R syntax! Do you know if ggplot will calculate the proportions for me first? Previously I had to use ddply to summarise my data before I plotted it?!?
I'd have to see what your original dataframe looks like, but what I usually do is calculate the proportion inside aes(). So, something like ggplot(data, aes(x, y=[value/sum(values)])). Anyway, is this what you're asking here?
Is there a general workaround? I'm using a variable to plot and hence can't subset the dataframe in before.
1

I'm unsure whether or not there is a way to set ignored values in ggplot. However you could consider simply recoding 0's to NA:

df[df$prop == 0] <- NA
Improve this answer

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.