ggplot2 - Stacking / dodging bar chart

问题:

I am having issues with making barchart in which the y axis is not count, but value from variables.

I use stat=identity, which is fine for one variable. But what if I have two variables and want to create a stacking / dodging barchart?

I have some mock data here:

foo <- data.frame(case=c('A','B','C'), var1=rnorm(3), var2=rnorm(3))

So my three cases on x-axis are A, B, and C. I want to plot the values of var1 and var2 as bars. Thanks!

回答1:

First, changed your sample data. When making data frame you don't need cbind() because in this case you make all columns as factors.

foo <- data.frame(case=c('A','B','C'), var1=rnorm(3), var2=rnorm(3))

To use two variables for barplot easiest way would be to melt the data from wide to long format.

library(reshape2)
foo.long<-melt(foo)
foo.long
  case variable      value
1    A     var1  0.7150827
2    B     var1 -0.5279363

Now use value as y values and variable for the fill. stat="identity" and position="dodge" will ensure that actual values are plotted and bars are dogged.

ggplot(foo.long, aes(case,value,fill=variable))+
      geom_bar(position="dodge",stat="identity")

回答2:

I am not answering the question, in the sens I am not using ggplot2. But, if you want to use the data.frame in its wide format, one other way is to use ``barchartfromlatticepackage. To getggplot2theme, I am usinglatticeExtra`. Using @Didzis data:

library(latticeExtra)
 barchart(var1+var2~case,data=foo,origin=0,
     par.settings = ggplot2like(),axis=axis.grid,auto.key=TRUE)