Matplotlib | Box-and-Whisker Plot. Display mean, median, outliers

You may want to grasp the trend or variation of your data at a glance, but are at a loss as to how to draw a box-and-whisker diagram or how to display its elements.

In this article, I will carefully explain how to display box-and-whisker plots and their means, medians, and outliers through Matplotlib’s boxplot function.

Acquire the skills to accurately understand the characteristics of your data.

After reading this article, you will be able to effectively visualize the trends and variations in your dataset using Matplotlib’s boxplot!

Please refer to the following article for a violin chart that adds data distribution to the box-and-whisker plot

あわせて読みたい

Matplotlib | Violin Plot. Mean, Median, Extrema, and Color Explained (violinplot) You may want to accurately understand the distribution and characteristics of your data and visually capture the extreme regions and medians, but you may be …

Axes.boxplot function

Box-and-Whisker Plots are drawn by passing an array as an argument to the Axes.boxplot function.

General box-and-whisker plot

For 1D arrays, one box-and-whisker plot is drawn.
For 2D arrays, multiple box-and-whisker plots are drawn.

In this article, the number of elements in a 2-dimensional array is 3 → [[], [], []]

The following tabs explain the code and flowchart

import matplotlib.pyplot as plt
import numpy as np

# step1 Fix the random numbers generated
np.random.seed(19680801)
# step2 Create data
all_data = [np.random.normal(0, std, 100) for std in range(7, 10)]
labels = ['x1', 'x2', 'x3']
# step3 Create graph frames
fig, ax = plt.subplots()

# step4 Plot a box-and-whisker plot
# General box-and-whisker plot
ax.boxplot(all_data, labels=labels)
ax.set_title('basic plot')

ax.set_xlabel('X label')
ax.set_ylabel('Y label')

plt.show()

Notched box-and-whisker plot (notch)

Displays a notched box-and-whisker plot with a box indentation.

The notches represent the confidence interval (CI) around the median.

Enter 1 or True for the second argument, notch, of the Axes.boxplot function

# step4 Plot a box-and-whisker plot
# notched
ax.boxplot(all_data, 1, labels=labels)

plt.show()

Outliers (sym, whis)

Box-and-whisker plot shows outliers.

The outliers can be customized with colors and symbols and can be hidden.

Color and shape of outlier symbols (sym=’gD’)

Customize the symbols for box-and-whisker outliers

Enter a string in the third argument, sym, of the Axes.boxplot function.

g is green and D is diamond shape

# step4 Plot a box-and-whisker plot
# Outlier Symbols
ax.boxplot(all_data, sym='gD', labels=labels)

plt.show()

Without the Outliers (sym=’empty string’)

Do not display outlier points in box-and-whisker plots

The third argument, sym, of the Axes.boxplot function is passed an empty string.

# step4 Plot a box-and-whisker plot
# Do not show outliers
ax.boxplot(all_data, sym='', labels=labels)

plt.show()

Range of outliers (whis)

Adjust the length of the box-and-whisker diagram to set the outlier range.

Enter a number in the fifth argument of the Axes.boxplot function, whis

1.5 is specified by default

# step4 Plot a box-and-whisker plot
# Adjust beard length to change the outlier region
ax.boxplot(all_data, sym='rs', whis=0.75, labels=labels)

plt.show()

Horizontal box-and-whisker plot (vert)

Displays a box-and-whisker diagram horizontally

Enter 0 or false for the fourth argument, vert, of the Axes.boxplot function

r is red, s is square shape

# step4 Plot a box-and-whisker plot
# Horizontal box-and-whisker plot
ax.boxplot(all_data, sym='rs', vert=False, labels=labels)

plt.show()

Color of box and frame (patch_artist)

Customize the color of the box on the box-and-whisker plot in a stylish way.

In the argument of the Axes.boxplot function, set patch_artist = True and replace it with the variable (bplot).

If patch_artist=True, the object becomes a Line2D object and can be customized.

Color of the box (set_color, set_facecolor)

The color of the boxes can be changed using set_color or set_facecolor in the [‘boxes’] attribute of the box-and-whisker plot.

Two arrays are used in a for statement simultaneously in the zip function

# step4 Plot a box-and-whisker plot
# Color of the box
bplot = ax.boxplot(all_data, labels=labels, patch_artist=True)
# Color List
colors = ['pink', 'lightblue', 'lightgreen']
# Assign colors to each box
for patch, color in zip(bplot['boxes'], colors):
    patch.set_color(color)

plt.show()

# step4 Plot a box-and-whisker plot
# Color of the box
bplot = ax.boxplot(all_data, labels=labels, patch_artist=True)
# Color list
colors = ['pink', 'lightblue', 'lightgreen']
# Assign colors to each box
for patch, color in zip(bplot['boxes'], colors):
    patch.set_facecolor(color)

plt.show()

Box Frame (edgecolor, linewidth)

You can assign a box color with set_edgecolor and a box thickness with set_linewidth to the [‘boxes’] attribute of the box-and-whisker plot.

# step4 Plot a box-and-whisker plot
# Color of the box
bplot = ax.boxplot(all_data, labels=labels, patch_artist=True)
# Color list
colors = ['pink', 'lightblue', 'lightgreen']
# Assign colors to each box
for patch, color in zip(bplot['boxes'], colors):
    # Box Color
    patch.set_color('white')
    # Frame Color
    patch.set_edgecolor(color)
    # Frame thickness
    patch.set_linewidth(3)

plt.show()

Average (showmeans, meanline)

In Matplotlib, the median is displayed by default for box-and-whisker plots, but the mean can also be displayed.

There are two ways of displaying the symbols and lines, and each will be explained separately.

Average with a symbol (showmeans)

Set showmeans=True in the Axes.boxplot function argument

# step4 Plot a box-and-whisker plot
# Average value
ax.boxplot(all_data, labels=labels, showmeans=True)

plt.show()

Average with a line (meanline)

Set showmeans=True and meanline=True as arguments to the Axes.boxplot function.

# step4 Plot a box-and-whisker plot
# Average value
ax.boxplot(all_data, labels=labels, showmeans=True, meanline=True)

plt.show()

Customize median and mean (medianprops, meanprops)

Feel free to tinker with the median and mean of the box-and-whisker diagram!

Set each parameter in dictionary type to medianprops and meanprops

Median color and line style (medianprops)

Enter a dictionary in the medianprops argument of the Axes.boxplot function.

# step4 Plot a box-and-whisker plot
# Median
ax.boxplot(
    all_data, 
    labels=labels, 
    medianprops={
        'color': 'C0',
        'linewidth':3,
        'linestyle': '-.',
    }
)

plt.show()

Average color and line style (meanprops)

Enter a dictionary in the meanprops argument of the Axes.boxplot function

Also, the median is hidden.

# step4 Plot a box-and-whisker plot
# Average
ax.boxplot(all_data, labels=labels, showmeans=True,
            # median not shown
            medianprops={
                'visible': False
            },
            # Customize averages
            meanprops={
                'marker': 'v',
                'markersize': 7,
                'markerfacecolor': 'white',
                'markeredgecolor': '#0097a7',
                'markeredgewidth': 2,
            }
)

plt.show()

References

Box-and-whisker diagram demonstration

Box-and-whisker diagram color settings