Bar charts are a very common and popular method of data visualization

Matplotlib makes it easy to plot bar charts with just a few lines of code.

This article describes how to plot general bar charts, stacked bar charts, grouped bar charts, and horizontal bar charts.

I also explained in detail how to label bar charts, set up error bars, and customize colors.

General bar chart (Axes.bar)

To plot a bar chart in Matplotlib, use the Axes.bar function

You can draw a bar graph with only the y height

The following tabs explain the code and flowchart

import matplotlib.pyplot as plt
import numpy as np
    
# step1 Create data
labels = ['G1', 'G2', 'G3', 'G4', 'G5']
x = np.arange(len(labels))
men_means = [20, 34, 30, 35, 27]
women_means = [25, 32, 34, 20, 25]

# step2 Create graph frames
fig, ax = plt.subplots()
# step3 Plot a bar chart
ax.bar(x, men_means, label='Men', tick_label=labels)

ax.set_xlabel('X label')
ax.set_ylabel('Y label')
ax.set_title('Basic Bar')
ax.legend()

plt.show()

Horizontal bar chart (Axes.barh)

To plot horizontal bar charts in Matplotlib, use the Axes.barh function

Basically, it is used in the same way as the Axes.bar function to draw a vertical bar chart.

# step3 Plot a horizontal bar chart
ax.barh(y, women_means, label='Women', tick_label=labels)

plt.show()

Set y-axis labels in ascending order (Axes.invert_yaxis)

The y-axis of the horizontal bar graph increases in value as you move up from the (0, 0) coordinate

However, when displaying with string labels, it is easier to see the opposite,
so Axes.invert_yaxis sets the y-axis labels in ascending order

# step3 Plot a horizontal bar chart
ax.barh(y, women_means, label='Women', tick_label=labels)
# Reverse the y-axis
ax.invert_yaxis()
plt.show()

Produce right-to-left bar charts (Axes.invert_xaxis)

The y-axis of the horizontal bar graph has bars from the x=0 coordinate

Axes.invert_xaxis produces a bar chart from right to left, with x=0 on the right

# step3 Plot a horizontal bar chart
ax.barh(y, women_means, label='Women', tick_label=labels)
# Reverse the x-axis
ax.invert_xaxis()
plt.show()

Stacked bar chart (bottom or left)

The Axes.bar and Axes.barh functions are also used to plot stacked bar charts

Define bottom as an argument and change the starting point of the second data and pile it up

The horizontal bar in the Axes.barh function defines left

# step3 Plot bar charts
ax.bar(x, men_means, label='Men')
ax.bar(x, women_means, label='Women', bottom=men_means, tick_label=labels)

plt.show()

Grouped bar chart (width or height)

The Axes.bar and Axes.barh functions are also used to group bars

To shift each element, define width

In this article, the width of the bars is set to 0.4 and the x-coordinate is shifted by 0.2

The horizontal bar chart in the Axes.barh function defines height

width = 0.4

# step3 Plot bar charts
ax.bar(x - width/2, men_means, width, label='Men', tick_label=labels)
ax.bar(x + width/2, women_means, width, label='Women', tick_label=labels)

plt.show()

Labeled bar chart (Axes.bar_label)

To display a label for each element of a bar chart, use the Axes.bar_label function

General bar chart with labels

This section explains how to label a single bar chart with the data as introduced in General bar chart

The procedure is to create a variable barcontainer with the Axes.bar function and define an array of labels in the variable.

# step3 Plot bar charts
# create a variable barcontainer
basic = ax.bar(x, men_means, width=0.4, label='Men', tick_label=labels)

# step4 Label bar charts
# define an array of labels
ax.bar_label(basic, labels=men_means)

plt.show()

Labeled stacked bar chart (fmt)

This section describes how to label a Stacked bar chart with multiple data in a vertical line

Each variable created by the Axes.bar function should have a different name

Also, fmt is now displayed to one decimal place.

width = 0.4

# step3 Plot bar charts
stock1 = ax.bar(labels, men_means, width, label='Men', tick_label=labels)
stock2 = ax.bar(labels, women_means, width, bottom=men_means, label='Women', tick_label=labels)

# step4 Label bar charts
ax.bar_label(stock1, fmt='%.1f')
ax.bar_label(stock2, fmt='%.1f')

plt.show()

Labeled grouped bar chart (padding)

This section describes how to label a Grouped bar chart with multiple data side by side.

Each variable created by the Axes.bar function should have a different name

Also, the spacing of the Women label has been changed by specifying padding

width = 0.4

# step3 Plot bar charts
group1 = ax.bar(x - width/2, men_means, width, label='Men', tick_label=labels)
group2 = ax.bar(x + width/2, women_means, width, label='Women', tick_label=labels)

# step4 Label bar charts
ax.bar_label(group1, labels=men_means)
ax.bar_label(group2, labels=women_means, padding=10)

plt.show()

Labeled horizontal bar chart (fontsize)

This section describes how to label Horizontal bar charts

fontsize is set to 15 to make the labels a little larger.
The others have fontsize=12.

# step3 Plot horizontal bar charts
graph = ax.barh(y, women_means, height=0.6, label='Women')
# step4 Label bar charts
ax.bar_label(graph, labels=women_means, padding=5, fontsize=14)

plt.show()

Error bars on bar charts (xerr, yerr)

The error bars of the bar chart define an array of numbers in xerr or yerr

xerr and yerr are the difference between x- and y-directions, and can be defined simultaneously.

Error bars in y direction (yerr)

The basic bar chart is an error bar in the y direction with yerr

# Array of error bars
men_std = [2, 3, 4, 1, 2]

# step3 Plot bar charts
ax.bar(x, men_means, width=0.4, label='Men', yerr=men_std, tick_label=labels)

plt.show()

Label the error bars in the bar chart. (xerr)

This section describes how to label the error bars on a horizontal bar chart to indicate the magnitude of the error.

In the horizontal bar chart, xerr is the error bar in the x direction

# Array of error bars
women_std = [3, 5, 2, 3, 3]

# step3 Plot horizontal bar charts
graph = ax.barh(y, women_means, height=0.6, label='Women', xerr=women_std)
# step4 Label bar charts
ax.bar_label(graph, labels=['±%.1f' % e for e in women_std], padding=5, fontsize=14)

plt.show()

Change the color of the error bars (ecolor)

The color of the error bar is defined by ecolor argument to the Axes.bar function.

Please refer to the following for how to specify colors

# Array of error bars
men_std = [2, 3, 4, 1, 2]

# step3 Plot bar charts
ax.bar(x, men_means, width=0.4, yerr=men_std, ecolor='red', tick_label=labels)

plt.show()

Change the size of the error bars (capsize)

The size of the error bar is defined by the capsize argument to the Axes.bar function

# Array of error bars
men_std = [2, 3, 4, 1, 2]
# step3 Plot bar charts
ax.bar(x, men_means, width=0.4, yerr=men_std, capsize=10, tick_label=labels)

plt.show()

Bar chart color and other customizations

The Axes.bar function has many other arguments that can be customized.

The color of the bars can be specified in two ways: color, the color of the surface of the bars, and edgecolor, the color of the border.

The thickness of the bar chart border can also be adjusted by combining edgecolor and linewidth

Change the main color of the bar (color)

The color of the bar surface is defined by the color argument to the Axes.bar function

This time, the strings C0, C1, and C2 are specified for each bar

# step3 Plot a bar chart
ax.bar(x, men_means, tick_label=labels, color=['C'+str(i) for i in range(len(men_means))])

plt.show()

Change the color of the bar edge (edgecolor, linewidth)

The color of the bar’s edges is defined by the edgecolor argument to the Axes.bar function.

This time, the strings C0, C1, and C2 are specified for each bar

The thickness of the border is adjusted numerically by linewidth, but it can also be specified as an array

# step3 Plot a bar chart
ax.bar(x, men_means, tick_label=labels, color='white', edgecolor=['C'+str(i) for i in range(len(men_means))], linewidth=5)

plt.show()

Align the x-coordinate bar (align)

Bars are drawn from the center of the specified x-coordinate

If align=edge, the x-coordinate is changed to the left end of the bar.

# step3 Plot a bar chart
ax.bar(x, men_means, align='edge', label='Men')

plt.show()

Bar chart on a logarithmic scale (log)

This section explains how to change the axis scale to logarithmic.

If log=True, the bar chart have a logarithmic axis.

# step3 Plot a bar chart
ax.bar(x, men_means, tick_label=labels, log=True)

plt.show()

References

Axes.bar function

Axes.barh function

Axes.bar_label function

Stacked bar chart

Grouped bar chart

Horizontal bar chart

Bar Chart Labels