R ggplot2: complex stacked barchart with multiple categorical variables

Question

My dataset in R looks like the following:

a <- c("M","F","F","F","M","M","F","F","F","M","F","F","M","M","F")
p <- c("P","P","W","W","P","P","W","W","W","W","P","P","P","W","W")
y1 <- c("yes","yes","null","no","no","no","yes","null","no","yes","yes","yes","null","no","no")
y2 <- c("yes","null","no","no","no","yes","yes","yes","null","no","yes","null","no","yes","yes")
y3 <- c("no","no","no","yes","null","yes","null","no","no","no","yes","yes","null","no","no")
VE <- data.frame(gender = a,
             type = p,
             y1 = y1,
             y2 = y2,
             y3 = y3)

And I would like to create a bar chart which looks like this: ideal bar chart

I just figured out a long way to get the chart:

q<-data.frame(gender=VE$gender,
          year=rep("y1",15),
          group=VE$y1)
p<-data.frame(gender=VE$gender,
          year=rep("y2",15),
          group=VE$y2)
x<-data.frame(gender=VE$gender,
          year=rep("y3",15),
          group=VE$y3)
Table<-rbind(q,p,x)
ggplot(Table, aes(year)) + geom_bar(aes(fill=group), position = "stack") + facet_grid(gender~.)

Is there any better way to get the bar chart? (since I was originally going to deal with 3,000,000 obsevations which have 32 variables each) Please give me some kind help with this bar chart. Cheers!

showing your data in an image does not make it easy to duplicate. Read [this](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) and the help and edit your question. Or search for your [question](https://stackoverflow.com/questions/38877580/stacked-bar-plot-with-4-categorical-variables-in-r?rq=1) It is in the related questions on the side of your questions after all. — shea, Oct 09 '17 at 21:11
Possible duplicate of [Stacked bar plot with 4 categorical variables in R](https://stackoverflow.com/questions/38877580/stacked-bar-plot-with-4-categorical-variables-in-r) — shea, Oct 09 '17 at 21:11

score 0 · Answer 1 · answered Oct 09 '17 at 21:58

First you can melt your data.frame to get a 'long' format. For this I have created an ID variable, the 3 variables 'y1, 'y2', and 'y3' are put together into one variable. You can then use ggplot2 and use geom_bar() which will count the values in the x aesthetic if no y aesthetic is provided.

library(ggplot2)

# create data frame
df <- data.frame(ID = 1:15, 
             gender = c('M', 'F', 'F', 'F', 'M', 'M', 'F', 'F', 'F', 'M', 'F', 'F', 'M', 'M', 'F'),
             type = toupper(c('p', 'p', 'w', 'w', 'p', 'p', 'w', 'w', 'w', 'w', 'p', 'p', 'p', 'W', 'W')),
             y1 = c('yes', 'yes', 'null', 'no', 'no', 'no', 'yes', 'null', 'no', 'yes', 'yes', 'yes', 'null', 'no', 'no'),
             y2 = c('yes', 'null', 'no', 'no', 'no', 'yes', 'yes', 'yes', 'null', 'no', 'yes', 'null', 'no', 'yes', 'yes'),
             y3 = c('no', 'no', 'no', 'yes', 'null', 'yes', 'null', 'no', 'no', 'no', 'yes', 'yes', 'null', 'no', 'no'),
             stringsAsFactors = TRUE)

# melt data frame to long format
df_melt <- data.table::melt(df[, c(1, 4:6)], id.vars = "ID")

# set correct levels for factor (needed for the legend)
df_melt$value <- factor(df_melt$value, levels = c("yes", "no", "null"))

# add ggplot
ggplot(data = df_melt) + 
  geom_bar(aes(x = variable, fill = value, colour = value)) +
  ylab("count") +
  xlab("year")

Which returns:

Basically, it's the melt() function in "reshape" package which automatically implements my long-way transform. Thanks! — J.wz, Oct 09 '17 at 22:20

R ggplot2: complex stacked barchart with multiple categorical variables

1 Answers1