Fill boxplot by variable but keep date scale GGplot

user_9144036 0
Viewed 30

I have a dataframe containing values for three variables at three different dates:

df <- data.frame(date = as.Date(c(rep('2021-01-01',6),rep('2021-05-04',6),rep('2015-04-04',6))),
             variable = rep(c('apple','bananna','carrot'),6),
             value = rnorm(18,5,2))

I'd like to make a boxplot with fill relating to variable, date on the x axis and value on the y axis:

ggplot(df,aes(x=as.factor(date),y=value,fill=variable))+geom_boxplot()

This fills based on variable, but it doesn't scale date according to the actual differences in time: enter image description here

If I set group=date, it gets scaled properly, but fill=variable stops working:

ggplot(df,aes(x=date,y=value,group=date,fill=variable))+geom_boxplot()

enter image description here

How do I get the date to scale properly and also fill based on variable name?

ggplot2 r
2 Answers

Seems like this was already answered here. You need to create an interaction for the group: group=interaction(variable,date):

   ggplot(df,aes(x=date,y=value,group=interaction(variable,date),fill=variable))+geom_boxplot()

enter image description here

user_91440360

The problem can be solved with the creation of a factor variable combining date and variable. This is done with interaction.

I have also added code to increase the plot window width, in order to make the bars and labels more visible.

df <- data.frame(date = as.Date(c(rep('2021-01-01',6),rep('2021-05-04',6),rep('2015-04-04',6))),
                 variable = rep(c('apple','bananna','carrot'),6),
                 value = rnorm(18,5,2))

library(ggplot2)

x11(width = 20, height = 8)    # this is not stricktly needed

ggplot(df, aes(x = date, y = value, fill = variable)) +
  geom_boxplot(aes(group = interaction(date, variable))) +
  scale_x_date(
    breaks = unique(df$date),
    date_labels = "%Y-%m-%d"
  ) +
  theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust = 1))

Created on 2022-06-02 by the reprex package (v2.0.1)

The same can be done by creating a temporary new factor in a dplyr pipe to mutate.

suppressPackageStartupMessages({
  library(ggplot2)
  library(dplyr)
})

df %>%
  mutate(group = interaction(date, variable)) %>%
  ggplot(aes(x = date, y = value, fill = variable)) +
  geom_boxplot(aes(group = group)) +
  scale_x_date(
    breaks = unique(df$date),
    date_labels = "%Y-%m-%d"
  ) +
  theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust = 1))

Created on 2022-06-02 by the reprex package (v2.0.1)

user_82454060
Related