Subplots, figures and axes (explicit references)

This page has 0 threads | Add post

Blogs

Finding our blogs useful? Please consider helping us to share more!

An introduction to using matplotlib in Python to create charts Part two of a six-part series of blogs
If you want to create charts in Python, the chances are that you'll do it using the matplotlib module. This blog will get you started and explain some of its foibles! Getting started with matplotlib Subplots, figures and axes (explicit references) (this blog) An introduction to formatting matplotlib charts Working with different chart types Saving charts Using style sheets to format charts consistently

An introduction to using matplotlib in Python to create charts

Part two of a six-part series of blogs

If you want to create charts in Python, the chances are that you'll do it using the matplotlib module. This blog will get you started and explain some of its foibles!

Getting started with matplotlib
Subplots, figures and axes (explicit references) (this blog)
An introduction to formatting matplotlib charts
Working with different chart types
Saving charts
Using style sheets to format charts consistently

Posted by Andy Brown on 02 June 2025

In this blog

The horrible matplotlib terminology
Creating single charts (including explaining unpacking)
Creating multiple charts

Nearly all Python programmers use explicit references to create objects for each chart they are building. Usually these objects consist of a single figure containing one or more subplots, each of which can contain multiple axes:

This figure contains a subplot consisting of one row and two (unequal) columns, and hence two axes (or charts).

The horrible matplotlib terminology

Here are the words that matplotlib uses to describe bits of charts like the one shown above. I've included the words they should have used too!

Word	What it describes	What it should have been called
Figure	The container for all of the subplots you're creating	Canvas
Subplot	A container for multiple charts	Subplot is fine!
Axes	A single chart within a subplot	Anything but this! Chart? Graph?

What were the authors of matplotlib thinking when they chose to call charts axes?

My award for the worst name-giver in computing would go to the person who gave the name Power Automate to two completely different Microsoft applicatons, but the person who called charts "axes" would be a close runner-up on my podium.

Creating single charts (including explaining unpacking)

Let's reproduce the chart from the previous part of this blog, but this time use an explicit reference to the chart (ahem, "axes") object:

The same chart using explicit references

We'll reproduce pretty much the same chart, as above.

Here's the code to do this:

import numpy as np

import matplotlib.pyplot as plt

# get the number of days of Xmas ( 1 to 12)

days_till_xmas = np.arange(1,13)

# panic levels increase exponentially

panic_level = days_till_xmas ** 2

# create a single chart within a subplot

fig, ax = plt.subplots(figsize = (5,3))

# customise this chart, now known as "ax"

ax.plot(days_till_xmas,panic_level)

# add a title

ax.set_title('Panic levels over the 12 days of Christmas')

# label the axes

ax.set_ylabel("Panic level")

ax.set_xlabel("Day number")

# space this out

fig.tight_layout()

# show the chart on screen

plt.show()

There are two things worth pointing out in this. The first is that we unpack the two things returned from subplots:

The subplots function returns a tuple containing the canvas containing your charts (the "figure") and a collection of the charts within it.

Our code would have been identical (if slighly longer) if we'd written this instead:

# create a single chart within a subplot

things_returned = plt.subplots(figsize = (5,3))

# get each of the two things returned

fig = things_returned[0]

ax = things_returned[1]

One very good reason not to spell out what youi're doing like this is that ... nobody else does.

The other thing to point out is that (as a general rule) you change from implicit to explicit formatting by adding set_ in front of formatting commands. Compare the implicit way:

# add a title

plt.title('Panic levels over the 12 days of Christmas')

# label the axes

plt.ylabel("Panic level")

plt.xlabel("Day number")

With the explicit way:

# add a title

ax.set_title('Panic levels over the 12 days of Christmas')

# label the axes

ax.set_ylabel("Panic level")

ax.set_xlabel("Day number")

So the functionality is the same, but the commands are slightly different.

Creating multiple charts

The great advantage of subplots is that they can contain more than one chart - like this, in fact:

A column chart taking up a wide column followed by a pie chart taking up the other (narrower) column.

You can use subplots to create multiple charts which share axes using the sharex and sharey keyword, although this blog doesn't cover this.

Here's how to create these charts:

import numpy as np

import matplotlib.pyplot as plt

# Wikipedia list of most populous countries in the world

countries = ["India","China","United States","Indonesia",

"Pakistan","Nigeria","Brazil"]

populations = [1413,1408,340,282,241,224,213]

# create a chart based on these countries amd figures

fig, (ax1,ax2) = plt.subplots(

nrows = 1,

ncols = 2,

figsize = (12,6),

gridspec_kw= {'width_ratios':[2,1]}

)

# create and format the bar chart on the left

ax1.bar(countries,populations)

ax1.set_title("Most populous countries")

ax1.set_xlabel("Countries")

ax1.set_ylabel("Populations (m)")

# create and format the pie chart on the right

ax2.pie(populations, labels = countries)

ax2.set_title("Countries by population")

# make things fit in better

fig.tight_layout()

plt.show()

The arguments to the subplots function above are as follows:

Argument	Value	What it denotes
nrows	1	There will be one row of charts
ncols	2	There will be two columns
figsize	(12,6)	The total canvas will be twice as wide as it is high
gridspec_kw	{'width_ratios':[2,1]}	A single-item dictionary setting the first column to be twice the width of the second column

If you're wondering how I found out about the gridspec_kw argument, it's the same way as you will: Googling, taking examples from the matplotlib website and using AI tools like ChatGPT!

Parts of this blog
Getting started with matplotlib Subplots, figures and axes (explicit references) (this blog) An introduction to formatting matplotlib charts Working with different chart types Saving charts Using style sheets to format charts consistently

Some other pages relevant to the above blogs include:

This blog has 0 threads

Add a new post

ALL BLOGS

PYTHON BLOGS

PYTHON BLOGS

The horrible matplotlib terminology

Creating single charts (including explaining unpacking)

Creating multiple charts

Head office

London

Manchester