Question

I have a dataframe with 3 columns

• $x — at http://pastebin.com/SGrRUJcA
• $y — at http://pastebin.com/fhn7A1rj
• $z — at http://pastebin.com/VmVvdHEE

that I wish to use to generate a stacked barplot. All of these columns hold integer data. The stacked barplot should have the levels along the x-axis and the data for each level along the y-axis. The stacks should then correspond to each of $x, $y and $z.

UPDATE: I now have the following:

counted <- data.frame(table(myDf$x),variable='x')
counted <- rbind(counted,data.frame(table(myDf$y),variable='y'))
counted <- rbind(counted,data.frame(table(myDf$z),variable='z'))
counted <- counted[counted$Var1!=0,]  # to get rid of 0th level??

stackedBp <- ggplot(counted,aes(x=Var1,y=Freq,fill=variable))
stackedBp <-  stackedBp+geom_bar(stat='identity')+scale_x_discrete('Levels')+scale_y_continuous('Frequency')
stackedBp

which generates:

.

Two issues remain:

1. the x-axis labeling is not correct. For some reason, it goes: 46, 47, 53, 54, 38, 40…. How can I order it naturally?

2. I also wish to remove the 0th label.

I’ve tried using +scale_x_discrete(breaks = 0:50, labels = 1:50) but this doesn’t work.

NB. axis labeling issue: Dataframe column appears incorrectly sorted

Answer 1

Not completely sure what you’re wanting to see… but reading ?barplot says the first argument, height must be a vector or matrix. So to fix your initial error:

myDf <- data.frame(x=sample(1:10,100,replace=T),y=sample(11:20,100,replace=T),z=1:10)
barplot(as.matrix(myDf))

If you provide a reproducible example and a more specific description of your desired output you can get a better answer.

Or if I were to guess wildly (and use ggplot)…

myDf <- data.frame(x=sample(1:10,100,replace=T),y=sample(11:20,100,replace=T),z=1:10)
myDf.counted<- data.frame(table(myDf$x),variable='x')
myDf.counted <- rbind(myDf.counted,data.frame(table(myDf$y),variable='y'))
myDf.counted <- rbind(myDf.counted,data.frame(table(myDf$z),variable='z'))

ggplot(myDf.counted,aes(x=Var1,y=Freq,fill=variable))+geom_bar(stat='identity')