Create Boxplot with respect to two factors using ggplot2 in R

Last Updated : 25 Nov, 2023

Multiple variable distributions can be visualized with boxplots. ggplot2 allows us to create beautiful boxplots quickly. It is possible to have multiple subgroups for a variable of interest. In those situations, it is very useful to visualize using “grouped boxplots”. The ggplot2 package in R programming language provides a number of options for visualizing such grouped boxplots.

Now talking about Boxplot, then each boxplot is created for every category or level in that factor that is represented by a factor and a numerical column. geom_boxplot also allows us to plot both factors using the fill argument when there are two factors. Geom_boxplot() is the key function

Syntax:

geom_boxplot(width,notch,color,size,linetype, fill,outliner.color, outliner.size, outliner.shape)

Parameter:

width: width of the boxplot

notch: if it is true then it will create a notched boxplot and notches are used to compare boxplots.

color, size,line type: borderline, color, size and shape.

fill: used to fill box plot areas.

outlier.colour, outlier.shape, outlier.size: The color, the shape and the size for outlying points.

Now let us look at a few implementations.

Example 1:

R

# create a Data Frame
Gender<-sample(c("Male","Female"),20,replace=TRUE)
Values<-rnorm(20,mean=0,sd=1)
Group<-sample(letters[1:5],20,replace=TRUE)
 
df<-data.frame(Gender,Values,Group)
 
library(ggplot2)
 
# creating a boxplot
ggplot(df,aes(Gender,Values))+geom_boxplot(aes(fill=Group))

Output:

Create a boxplot with respect to two factors using ggplot2

Example 2:

R

# load ggplot2 package if already installed
library(ggplot2)
 
# create a data frame with two factors 
df <- data.frame(Factor1=factor(rbinom(30, 1, 0.55), 
                                label=c("male","female")), 
                   Factor2=factor(rbinom(30, 1, 0.45), 
                                  label=c("young","old")),
                   Values=rnorm(30,mean=5,sd=2))
 
# Now make a interaction between two factors 
# on x axis  
df$Factor1Factor2 <- interaction(df$Factor1, df$Factor2) 
 
# now Plot Boxplot with fill color according
# to factor1 and factor2
ggplot(aes(y = Values, x = Factor1Factor2), data = df) + 
geom_boxplot(aes(fill=Factor1Factor2))
  

Output:

Create Boxplot with respect to two factors using ggplot2

Example 3:

R

# Load ggplot2 package if already installed
library(ggplot2)
 
# Create a data frame with two factors
df <- data.frame(
  Factor1 = factor(rbinom(30, 1, 0.55), label = c("male", "female")),
  Factor2 = factor(rbinom(30, 1, 0.45), label = c("young", "old")),
  Values = rnorm(30, mean = 5, sd = 2)
)
 
# Create an interaction between the two factors
df$Factor1Factor2 <- interaction(df$Factor1, df$Factor2)
 
# Define custom colors for the fill
custom_colors <- c("steelblue", "darkorange", "forestgreen", "firebrick")
 
# Plot the box plot with custom aesthetics
ggplot(df, aes(x = Factor1Factor2, y = Values, fill = Factor1Factor2)) +
  geom_boxplot(width = 0.5, alpha = 0.7, outlier.shape = NA) +
  geom_jitter(width = 0.2, height = 0, size = 3, alpha = 0.8) +
  scale_fill_manual(values = custom_colors) +
  labs(x = "Factor 1 & Factor 2", y = "Values") +
  ggtitle("Box Plot with Factor 1 & Factor 2") +
  theme_minimal() +
  theme(
    plot.title = element_text(size = 16, face = "bold"),
    axis.text = element_text(size = 12),
    axis.title = element_text(size = 14, face = "bold"),
    legend.title = element_blank(),
    legend.position = "none"
  )

Output:

Create boxplot with respect to two factors using ggplot2

Custom colors (custom_colors) use to establish the boxes’ fill. we can change these colors to suit our tastes or go with a whole other palette.
To overlay individual data points on top of the boxes, geom_jitter was added. This gives an idea of how the data is distributed.
Utilizing the geom_boxplot options, the boxes’ width, alpha (transparency), and outlier form were changed.
added more plot components, such as axis labels and a plot title, and changed the theme to a simple design.