Barplot visualization for a survey question based on checkboxes

Question

Hi!

I have to make a report of a survey where we have a question like this (all answers are checkboxes).

What is you favorite cake(s) (please choose more than one):
[] Tiramisù
[] Carrot Cake
[] Cupcake

Then the survey software exports into CSV like this:

"username","likes_tiramisu","likes_carrotcake","likes_cupcake"
"test01",1,1,1
"test02",0,1,1
"test03",0,1,0
"test04",0,0,1

I would like to make a barplot where every histogram represent the frequency of every cake. How can I combine a table like this:

"likes_tiramisu"  "likes_carrotcake"  "likes_cupcake"
               1                   3                3

Could I just solve it with sum of all elements in each column? Is it conceptually valid?

Answer 1

Suppose your data is in the file: survey.csv

Then to create a barplot use the following commands:

#Read in the data
d = read.table("survey.csv", sep=",", header=TRUE)

#Need to skip the username column, so d[,2:4]
#Use apply to calculate the totals in your table
barplot(apply(d[,2:4], 2, sum))

HTH

Answer 2

A ggplot2 approach would look like this:

library(ggplot2)

data.melt <- melt(data, id = "username")
qplot(variable, value, data = data.melt, geom = "bar", stat = "identity")