plotting two variables on bar graph using ggplot2

Question

I am new to using ggplot2 and i am having trouble plotting a graph. I have looked around on SO but the solutions I found did not work with my data. Here is an example of my DF

Count1     Count2      Color 
  3         4          Red
  3         6          Green 
  5         2          Red
  2         0          Blue

I would like to just plot this is as a bar graph. I would like the X axis to consist of the colors and I would like to plot both the Count1 and Count2 variables on the y axis. for example, the two bars used to show the green color will go up to the number 3 (for count1) and the number 6 (for count2). Similarly, the red bar will go up to 8 (for count1) and 6 (for count2) Does anyone know how to go about doing this? Thanks!

Aggregate and reshape your data first, e.g. library(tidyverse); df %>% group_by(Color) %>% summarise_all(sum) %>% gather(var, val, -Color) %>% ggplot(aes(x = Color, y = val, fill = var)) + geom_col(position = 'dodge') — alistaire
– alistaire, Commented Apr 19, 2017 at 0:51
@alistaire when I tried the summarise_all part of your code on my real data, I got this error message and im not sure what to make of it: Error: invalid 'type' (character) of argument — Amanda
– Amanda, Commented Apr 19, 2017 at 18:15
summarise_all tries to apply its function (sum here) to all non-grouping columns, which for the sample data is fine, but may not be for your real data if there are non-numeric columns. (Also use str to make sure your numeric columns are actually stored as doubles or integers.) summarise_at lets you specify columns, or just use summarise and explicitly specify what to do with each column you want. — alistaire
– alistaire, Commented Apr 19, 2017 at 21:28

Jeremy Voisey · Accepted Answer · 2017-04-19 05:34:14Z

2

Breaking up the answer from @alistaire above, so you can follow what's going on

Your data

color_df <- data.frame(Count1 = c(3,3,5,2), Count2 = c(4,6,2,0), Color = c("Red", "Green", "Red", "Blue"))

Adding up counts for each color

library(dplyr)
sum_df <- color_df %>%
    group_by(Color) %>%
    summarise_all(sum)
sum_df

ggplot needs both counts in one column, with another column describing which is which. Compare sum_df and tidy_df

library(tidyr)
tidy_df <- sum_df %>%
    gather(CountName, Count, -Color)

Finally the plot. Dodge puts them side by side. geom_col uses heights from Count variable. geom_bar would count number of observations itself.

library(ggplot2)
ggplot(tidy_df, aes(x = Color, fill = CountName, y = Count)) +
    geom_col(position = "dodge")

answered Apr 19, 2017 at 5:34

Jeremy Voisey

1,29710 silver badges14 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Amanda Over a year ago

Thanks Jeremy. However, when I tried the group_by and summarise_all part of your code on my real data, I got this error message and im not sure what to make of it: Error: invalid 'type' (character) of argument

Jeremy Voisey Over a year ago

Check your counts are stored as numbers.

Jeremy Voisey Over a year ago

The likely problem is the sum. Do you have other columns apart from the counts? If so trying to sum a character column would give this error. Try replacing summarise_all with summarise (Count1 = sum(Count1), Count2 = sum(Count2))

Collectives™ on Stack Overflow

plotting two variables on bar graph using ggplot2

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related