CODEWITHSUNDEEP
rggplot2
slhck
slhck

Reputation: 38740

Creating a bar plot from summarySE result table with multiple group variables

I have a result table from the summarySE function:

    a        b N               variable   sd    se   ci
 1234      foo 264                  2.0 0.87 0.053 0.11
 1234      bar 111                  3.6 1.35 0.128 0.25
 5678      foo 169                  1.8 1.02 0.079 0.16
 5678      bar 118                  1.6 1.13 0.104 0.21
91011      foo   9                  1.3 1.35 0.450 1.04
91011      bar 384                  1.0 1.12 0.057 0.11

I want to create a bar plot, where each row corresponds to a bar, its height being variable – so I need stat="identity". Now, usually, I would have no problem doing this:

column = "varaible"
ggplot(data, aes_string(y = column)) + geom_bar(stat = "identity")

But it fails with:

Error during wrapup: argument "env" is missing, with no default

Of course, because there are multiple columns defining what x is. If I do

ggplot(data, aes_string(x = "a", y = column)) + geom_bar(stat = "identity")

It works for unique values of a, but it does not consider b. How do I consider both combinations of a and b as x values?

Upvotes: 0

Views: 242

Answers (1)

lukeA
lukeA

Reputation: 54247

You could use another aesthetic like fill 

ggplot(data, aes_string(x = "a", fill = "b", y = column)) + 
  geom_bar(stat = "identity")

or concatenate both columns 

ggplot(transform(data, t = paste(a, b)), 
       aes_string(x = "t", y = column)) + 
  geom_bar(stat = "identity")

Upvotes: 1

Related Questions