Giving data frame like this:
sam<-data.table(title=c(rep("Cat",8),rep("Dog",4)),
fcat=c("A","B","C","B","B","C","C","B","C","B","B","C"),
fnum=c(seq(8,1),seq(4,1)),
labeltext=c("Pancakes","Muffins","Baursaq","Muffins","Muffins",
rep("Baursaq",3),"Muffins","Baursaq","Baursaq","Muffins"),
size=c(10,rep(1,11)))
sam
title fcat fnum labeltext size
1: Cat A 8 Pancakes 10
2: Cat B 7 Muffins 1
3: Cat C 6 Baursaq 1
4: Cat B 5 Muffins 1
5: Cat B 4 Muffins 1
6: Cat C 3 Baursaq 1
7: Cat C 2 Baursaq 1
8: Cat B 1 Baursaq 1
9: Dog C 4 Muffins 1
10: Dog B 3 Baursaq 1
11: Dog B 2 Baursaq 1
12: Dog C 1 Muffins 1
I want to create a barplot with x=title, y=size, filled by fcat, labelled by labeltext and ordered in descending order by fnum. First approach without reordering the original data. Here I concatenated text of labels in order to check if the order, size and category is used correctly.
ggplot(sam)+
geom_bar(aes(x=factor(title), y=size,fill=fcat),
stat="identity", color="white", size=1)+
geom_text(aes(x=title, y=size,
label=(paste(labeltext,size,fnum,fcat, sep=" "))),
size = 3, position = position_stack(vjust = 0.5))+
ggtitle("Without ordering")
Labels are ordered as needed, but there's a big problem with filling. Reorder filling by fnum:
ggplot(sam)+
geom_bar(aes(x=factor(title),
y=size,fill=reorder(fcat,fnum)),
stat="identity", color="white", size=1)+
geom_text(aes(x=title,y=size,
label=reorder(paste(labeltext,size,fnum,fcat,
sep=" "),fnum)),
size = 3, position = position_stack(vjust = 0.5))+
ggtitle("Ordered by fnum")
In this case bars are grouped by wrong fillings. For instance, for "Cat" the graph should plot 1 blue sized 10 bar, then 1 green, 1 red, 2 greens, 2 red, 1 green (as in labeltext's categories).
I want the plot to be stripped as in the original data. Like this guy had in first place. stat="identity"
doesn't work as expected.
I tried numerous approach reordering, but nothing helps much. I've searched through the SO barplot fill questions in ggplot2 tag (more than 700 posts o_O) and read the documentation very carefully. Nonetheless, no answer was found so far.