R : How to plot two different kind of plot with ggplot2?

Question

My data is like this :

id <- c('a','a','b','b')
var <- c('Var1','Var2','Var1','Var2')
value <- c(123,421,2132,633)
df <- cbind(id,var,value)

I would like to plot a scatter plot of Var1 and a bar chart of Var2 , with a legend showing that the dot is Var1 and the Bar is Var2

I have tried these code :

g <- ggplot(data = df, aes(x=id, y=value, fill=var)
g + stat_identity(geom = c("point","bar"))

But both Var1 and Var2 are showed in scatter point in this way.

I also tried to create two seprate data for Var1 and Var2 .

id = c('a','b')
Var1 = c(123,2132)
Var2 = c(421,633)
df1 = cbind.data.frame(id, Var1)
df2 = cbind.data.frame(id, Var2)

g <- ggplot()
g + geom_point(data = df1, aes(x=id, y=Var1),stat = "identity")) +
    geom_bar(data = df2, aes(x=id, y=Var2),stat = "identity"))

In this way, I can combine a scatter point with a bar chart, but there is no legend and I don't know how to make the legend telling readers that the dot is Var1 and the bar is Var2

Answer 1

First off, see this post . Thanks @Daniel!

You could try this:

id = c('a','b')
Var1 = c(123,2132)
Var2 = c(421,633)
df1 = cbind.data.frame(id, Var1)
df2 = cbind.data.frame(id, Var2)
library(ggplot2)
g <- ggplot()
g + geom_bar(data = df2, aes(x=id, y=Var2, fill = 'Var2'),stat = "identity")+
     geom_point(data = df1, aes(x=id, y=Var1, colour = 'Var1'),stat = "identity", size = 2) +
    scale_fill_manual(values = c('Var2' = 'blue'), name = 'Bars')+
    scale_colour_manual(values = c('Var1' = 'red'), name = 'Dots')

Basically, we use the aes within geom_point and geom_bar to define our color or fill. Then use scale_*_manual to control display. Also, we call geom_bar before geom_point so that the dots are plotted over the bar. Also, I use size = 2 in geom_point so that the point is a little more distinguishable.