R: how to turn variables to values in dataframe for ggplot

Question

I want to create a barplot that shows the numbers of bugs in 3 different lifestages. For each date (x) I want three bars representing the nr. of individuals in each of the lifestages.

My dataframe from the raw data looks similar like this (simplified example):

# create dataframe
date <- c("01/02/2018","14/02/2018","20/02/2018","03/03/2018","15/03/2018")
adult <- c(5,2,3,1,1) 
larvae <- c(6,5,9,7,12) 
nymph <- c(4,4,8,13,10)
df <- data.frame(date,adult,larvae,nymph)

 date adult larvae nymph
1 01/02/2018     5      6     4
2 14/02/2018     2      5     4
3 20/02/2018     3      9     8
4 03/03/2018     1      7    13
5 15/03/2018     1     12    10

The only way I know how to plot this with ggplot, is to turn the variables into factor levels of a new variable, say lifestage, and put all the counts into a variable counts.

That dataframe would look like this:

df2
date  stage counts
1  01/02/2018  adult      5
2  14/02/2018  adult      2
3  20/02/2018  adult      3
4  03/03/2018  adult      1
5  15/03/2018  adult      1
6  01/02/2018 larvae      6
7  14/02/2018 larvae      5
8  20/02/2018 larvae      9
9  03/03/2018 larvae      7
10 15/03/2018 larvae     12
11 01/02/2018  nymph      4
12 14/02/2018  nymph      4
13 20/02/2018  nymph      8
14 03/03/2018  nymph     13
15 15/03/2018  nymph      1

Plotting this df is easy:

ggplot(df2, aes(date, counts, fill=stage)) +
  geom_col(position = "dodge")

To get from df to df2 I have rather large workarounds involving extracting columns, creating new vectors with rep("stagename",x) to add to the dataframe, rbind() the whole dataframe times the nr of variables I want to turn into factor levels, etc. (I noticed I have used several methods before, but all quite long).

So I have 2 questions:

1) Is there a quick way to turn the different variables into factor levels of one new variable? I'm talking about a large dataframe with several other variables as well that need to stay.

2) Is there a way to get the same type of barplot without having to transfrom the dataframe?

I was trying something like this, but that's certainly not correct:

ggplot(data=df) +
  geom_col(aes(x=date,y=adult),fill="blue") +
  geom_col(aes(x=date,y=nymph),fill="green") +
  geom_col(aes(x=date,y=larvae),fill="yellow")

I've searched for similar questions, but can't seem to find a problem quite like mine. Also mine is double: if I can make the ggplot without the transformation, it would be better.

I've recently discovered the package tidyverse and assumed a solution for the transformation would lie in there, but I haven't come across anything yet that allows me a quick transformation of this kind. I'd prefer any solutions using that package if possible.

erocoar · Accepted Answer · 2018-08-24 12:09:26Z

4

library(tidyverse)
df %>% 
  gather(stage, counts, -date) %>%
  ggplot(aes(date, counts, fill = stage)) +
  geom_col(position = "dodge")

This should do it, or not?

answered Aug 24, 2018 at 12:09

erocoar

5,9333 gold badges26 silver badges45 bronze badges

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

2 Comments

Tingolfin Over a year ago

It does and it is fantastically short, love it! But how does it work? How does R (or gather) know what I mean with "stage" and "counts"? I'd like to understand it so I can play with it myself the next time. The help page does not explain how R links the key and value with the correct column.

erocoar Over a year ago

Hey, maybe it would have been more intuitive using gather("adult", "larvae", "nymph", key = "stage", value = "counts"). You select the columns that represent values, rather than variables. The key is the column names, which are called stage in your data, and the value in your data is counts. In my answer, gather() will gather all columns except date. -date (i.e. exclude date) comes from the dplyr::select() notation. This chapter might be of use to you!

Collectives™ on Stack Overflow

R: how to turn variables to values in dataframe for ggplot

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related