How to plot two columns in a stacked barplot

Question

I would appreciate your help in the following: I'm trying to build a stacked barplot. The fill is the split of business (new and renewed), the Y axis is volume of new/renewed business and the X would be the month. In my data I have two years of history, so I would like to have two stacked columns per month, one for n-1 year and one for n year. However, I don't know how to do this last step...

To clarify, please find below a picture of the data, the plot I have so far and a drawing of my goal.

And here is the code:

ggplot(BUSINESS, aes(fill=Business, y=GWP_mio, x=Date)) + 
  geom_bar(position="stack", stat="identity") +
  scale_fill_manual(values = c("#009E73","Darkblue")) +
  scale_y_continuous(breaks=seq(0,18000000,1000000), labels = scales::comma_format(scale = 1/1000000, 
  accuracy = .1), limits = c(0,18000000)) + 
  theme_ipsum() +
  ggtitle('GWP development') +
  theme(plot.title = element_text(hjust=0.5, size=14, family="Calibri", face="bold"),
        legend.title = element_text(size=11, family="Calibri", face="bold"),
        axis.title.x = element_text(hjust=1, size=11, family="Calibri", face="bold"),
        axis.text.x = element_text(angle=90, hjust=1, size=11, family="Calibri"),
        axis.title.y = element_text(angle=0, hjust=1, size=10, family="Calibri", face="bold"))

Any help would be highly appreciated.

[2 []3

Answer 1

What you're asking for is both position= 'dodge' and position= 'stack'. I'd actually suggest using faceting:

Data

library(data.table)
library(ggplot2)
N <- 24
dat <- data.table(
  date= rep(seq.Date(as.Date('2017-01-01'), as.Date('2018-12-01'), '1 month'), each= 2)
  , new= rpois(n= 2 * N, lambda= 5)
  , renew= rpois(n= 2 * N, lambda= 4)
)

dat[,year := data.table::year(date)]
dat[,month:= data.table::month(date)]
dat <- melt(dat, id.vars= c("date", "year", "month"), variable.name= 'business_type', value.name= 'units')

Facet

This is going to be much easier for the viewer.

ggplot(dat, aes(x= month, y= units, fill= factor(year))) +
  geom_bar(position= 'dodge', stat='identity') + facet_grid(business_type ~ .) +
  theme(axis.text.x= element_text(angle= 90))

Solution

But that's not what you asked for. So let's do something hacky. You'll have to mess around with the colours / fill to get exactly what you want. But here we add a first layer for the total, then a second layer for the new

N <- 24
dat <- data.table(
  date= rep(seq.Date(as.Date('2017-01-01'), as.Date('2018-12-01'), '1 month'), each= 2)
  , new= rpois(n= 2 * N, lambda= 5)
  , renew= rpois(n= 2 * N, lambda= 4)
)
dat[,year := data.table::year(date)]
dat[,month:= data.table::month(date)]
dat[, total := new + renew]

ggplot(dat, aes(x= month, y= total, fill= factor(year))) + 
  geom_bar(stat= 'identity', position= 'dodge', ) +
  geom_bar(data= dat, aes(x= month, y= new, fill= "black", colour= factor(year)), stat= 'identity', position= 'dodge') +
  scale_color_brewer(palette = "Dark2")