Johann Horvat Johann Horvat - 1 month ago 14
R Question

Overlay two geom_bar like two barplots with par(new=TRUE)

I currently have a plot similar to this:

sourceTable<-"dateValues;total;inHospital
2014-01-01; 80; 20
2014-01-02; 90; 15
2014-01-03; 98; 16
2014-01-04; 98; 17
2014-01-05; 100; 28
2014-01-06; 110; 28
2014-01-07; 122; 30
2014-01-08; 110; 25
2014-01-09; 100; 22
2014-01-00; 90; 20
2014-01-11; 90; 15
2014-01-12; 80; 13
2014-01-13; 82; 10"

timetable<-read.table(textConnection(sourceTable), sep=";", head=T, dec=",")
barplot(timetable$total, border="red", axes=FALSE, ylim=c(0, max(timetable$total)))
par(new=TRUE)
barplot(timetable$inHospital, border="blue", axes=FALSE, ylim=c(0, max(timetable$total)))


which gives me this plot:

two overlayed barplots

Is there a way to do this with ggplot and geom_bar like this:

library(ggplot2)
ggplot(timetable, aes(x=dateValues, y=total))
+geom_bar(stat="identity", fill="red", colour="red")


How can I overlay a second ggplot over the first one, i.e.:

ggplot(timetable, aes(x=dateValues, y=inHospital))
+geom_bar(stat="identity", fill="red", colour="blue")

Answer

Try this:

library(ggplot2)

ggplot(data = timetable, aes(x = dateValues, y = total)) + 
  geom_bar(stat = "identity", fill = "grey", colour = "red")+
  geom_bar(data = timetable, aes(x = dateValues, y = inHospital),
           stat = "identity", fill = "grey", colour = "blue")

enter image description here


EDIT: Alternative - the right way of doing it - transform the data before plotting:

library(reshape2)
library(ggplot2)

# transform the data - melt
timetable$outHospital <- timetable$total - timetable$inHospital
df <- melt(timetable, id = c("dateValues", "total"))

# plot in one go
ggplot(data = df, aes(x = dateValues, y = value, fill = variable)) + 
  geom_bar(stat = "identity")

enter image description here