Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

ggplot2 order categorical stacked bars by proportions of y-axis

I have a data frame with categorical x-axis called Category and the yaxis is the Abundance, colored by Sequence. For each Category I am trying to reorder the stacks by the Abundance, so that it is easily visualized which sequence has the highest proportion at the bottom, to the lowest proportion at the top.

Currently, I can make a bar graph like this:

s<-"Sequence Abundance Category
CAGTG 0.8 A
CAGTG 0.2 B
CAGTG 0.6 C
CAGTG 0.3 D
CAGTG 0.1 E
GGGAC 0.1 A
GGGAC 0.1 B
GGGAC 0.3 C
GGGAC 0.6 D
GGGAC 0.1 E
CTTGA 0.1 A
CTTGA 0.7 B
CTTGA 0.1 C
CTTGA 0.1 D
CTTGA 0.8 E"

d<-read.delim(textConnection(s),header=T,sep=" ")

g = ggplot(d,aes(x = Category, y = Abundance, fill = Sequence)) + 
      geom_bar(position = "fill",stat = "identity")

My data is very similar to this: Ordering stacks by size in a ggplot2 stacked bar graph

But even trying to reproduce this solution (following the steps in the answer), it does not reorder the stacks by proportion:

d$Sequence <- reorder(d$Sequence, d$Abundance)
d$Sequence <- factor(d$Sequence, levels=rev(levels(d$Sequence)))
ggplot(d, aes(x=Category, y=Abundance, fill=Sequence)) + 
  geom_bar(stat='identity') 

I cannot find an example for what I am looking for. Thanks so much for any help!

like image 752
user971102 Avatar asked Mar 08 '23 08:03

user971102


1 Answers

Use the group aesthetic to control the order of the stacked bar.

s <- "Sequence Abundance Category
CAGTG 0.8 A
CAGTG 0.2 B
CAGTG 0.6 C
CAGTG 0.3 D
CAGTG 0.1 E
GGGAC 0.1 A
GGGAC 0.1 B
GGGAC 0.3 C
GGGAC 0.6 D
GGGAC 0.1 E
CTTGA 0.1 A
CTTGA 0.7 B
CTTGA 0.1 C
CTTGA 0.1 D
CTTGA 0.8 E"  
d <- read.delim(textConnection(s), header=T, sep=" ")

# Add the "group" aesthetic to control the order of the stacked bars
g = ggplot(d,aes(x=Category, y=Abundance, fill=Sequence, group=Abundance)) + 
    geom_bar(position = "fill",stat = "identity")
g

enter image description here

like image 80
Marco Sandri Avatar answered Mar 23 '23 22:03

Marco Sandri