5
votes

I have been looking for a way to be able to show the stacked bar chart responses as percentage values, according to the gender classification of the respondents.

I was successful in creating a stacked bar plot using the variable 'sex' for the fill, but I want the plot to show the proportion between this variable. I know that using (..count..)/sum(..count) and scale_y can change the y axis so that it shows the percentages, but I can't find a way to use it for what I want. Manually doing a separate data frame with the frequency values [edit] reflecting the percentages is also possible, but I am really keen on looking for a way to use only ggplot.

This is what the plot currently looks like:

This is the current code:

workday<-ggplot(student,aes(x=Dalc2,fill=sex))
plot1<-workday+geom_bar()+facet_wrap(~romantic2)+labs(title="Workday Alcohol Consumption",y="Number of Respondents",x="Response")+ylim(0,300)

I know it's a pretty basic problem, but any enlightenment on this one would be greatly appreciated.

(Data set from uci.edu)

EDIT:

For those interested in the solution, r.bot (lots of thanks!) suggested using geom_bar(position="fill")

Also adding this to modify the y-axis: scale_y_continuous(labels=percent)

This is what the final graph looks like

1

1 Answers

10
votes

The ggplot online documentation has a nice example of stacked bars to 100%. I can't adapt your code as you haven't provided the data. As an example, see

# http://docs.ggplot2.org/current/geom_bar.html
require(ggplot2)

g <- ggplot(mpg, aes(class))
g + geom_bar(aes(fill = drv), position = "fill")

The position = "fill"should get you what you want.