ggplot geom_bar where x = multiple columns

Question

How can I go about making a bar plot where the X comes from multiple values of a data frame?

Fake data:

data <- data.frame(col1 = rep(c(\"A\", \"         


        

Answer 1

A very rough approximation, hoping it'll spark conversation and/or give enough to start.

Your data is too small to do much, so I'll extend it.

set.seed(2)
n <- 100
d <- data.frame(
  cat1 = sample(c('A','B','C'), size=n, replace=TRUE),
  cat2 = sample(c(2012L,2013L,2014L,2015L), size=n, replace=TRUE),
  cat3 = sample(c('^','v','<','>'), size=n, replace=TRUE),
  val = sample(c('X','Y'), size=n, replace=TRUE)
)

I'm using dplyr and tidyr here to reshape the data a little:

library(ggplot2)
library(dplyr)
library(tidyr)

d %>%
  tidyr::gather(cattype, cat, -val) %>%
  filter(val=="Y") %>%
  head
# Warning: attributes are not identical across measure variables; they will be dropped
#   val cattype cat
# 1   Y    cat1   A
# 2   Y    cat1   A
# 3   Y    cat1   C
# 4   Y    cat1   C
# 5   Y    cat1   B
# 6   Y    cat1   C

The next trick is to facet it correctly:

d %>%
  tidyr::gather(cattype, cat, -val) %>%
  filter(val=="Y") %>%
  ggplot(aes(val, fill=cattype)) +
  geom_bar() +
  facet_wrap(~cattype+cat, nrow=1)