CODEWITHSUNDEEP
rgroupingbar-chart
Fizza Tauqeer
Fizza Tauqeer

Reputation: 55

Creating Grouped Bar-plots for Multiple Cluster Columns in R

Desired Behavior: I want to create a grouped barplot of multiple columns of my dataset, plotting the grouped barplot with respect to a categorical variable having the values low, medium and high. For example, for the dataset shown below, I want to create a a three-group bar-plot for each 'B', 'M', 'LA' and so on on the x-axis, while filling the bars of each three-group by the mean of their values according to the category.

I have tried the following code to counter this, but I can't seem to understand what I am doing wrong here, as the code throws the error that the values don't add up dimension wise. Can anyone let me know what is the fault here, as per to my knowledge all the filled aspects of the ggplot seem to emulate my requirement? ...

df <- data.frame(B=c(1,2,4,-5,8,9,8,4,5),GM=c(5,5,4,-7,9,8,7,-4,6),
             LA=c(-5,8,5,-8,4,5,6,1,7), NS= c(-5,-8,-5,-8,4,5,-6,1,7), 
SL=c(-5,8,5,-8,-4,-5,6,-1,-7), data.SWASH_group=c("low", "medium", "low", 
"high", "high", "low", "medium", "low", "high"))

species=c(df$B, df$GM, df$LA, df$NS, df$SL)

condition=c(df$data.SWASH_group=="low", df$data.SWASH_group=="medium", 
df$data.SWASH_group=="high")

value=c(mean(df$B), mean(df$GM), mean(df$LA), mean(df$NS), mean(df$SL))

# For Grouping
library(ggplot2)

ggplot(df, aes(fill=condition, y=value, x=species)) + 
geom_bar(position="dodge", stat="identity")

This is how the head of my real dataset - df - looks like (The one in the 
code is just for example purposes):

B           GM          LA          NS          SL          data.SWASH_group
561.5806    533.5484    602.4000    675.4706    621.5758    low
780.5758    615.8788    801.8788    589.2903    875.0938    medium
649.9697    658.9091    789.1818    638.4412    646.5455    medium

... I expect the output to look something like this:

enter image description here

This is the output my own code produces:

enter image description here

Upvotes: 2

Views: 2315

Answers (1)

Dominix
Dominix

Reputation: 441

We would need reproducible data to address your question. But if I have understood your description correctly this should work:

library(ggplot2)
library(dplyr)
library(reshape2)

df <- data.frame(A=c(1,2,4,5,8,9,8,4,5),B=c(5,5,4,7,9,8,7,4,6),
                 C=c(5,8,5,8,4,5,6,1,7), 
                 group.var=c("low", "medium", "low", "high", "high", "low", "medium", "low", "high"))

df.agg <- df %>% group_by(group.var) %>% summarise_all(funs(mean))
df.agg <- melt(df.agg)

ggplot(df.agg) + 
  geom_bar(aes(x=variable, y = value, fill = as.factor(group.var)),
           position="dodge", stat="identity")

Upvotes: 1

Related Questions