matplotlib / cheatsheets

Official Matplotlib cheat sheets

Cheatsheets for Matplotlib users

Cheatsheets

Cheatsheet (download pdf)

Handouts

Beginner handout (download pdf)	Intermediate handout (download pdf)	Tips handout (download pdf)

For contributors to the cheatsheets

How to compile

1. You need to create a fonts repository with:

• fonts/roboto/* : See https://fonts.google.com/specimen/Roboto or https://github.com/googlefonts/roboto/tree/master/src/hinted
• fonts/roboto-slab/* : See https://fonts.google.com/specimen/Roboto+Slab or https://github.com/googlefonts/robotoslab/tree/master/fonts/static
• fonts/source-code-pro/* : See https://fonts.google.com/specimen/Source+Code+Pro or https://github.com/adobe-fonts/source-code-pro/tree/release/OTF
• fonts/source-sans-pro/* : See https://fonts.google.com/specimen/Source+Sans+Pro or https://github.com/adobe-fonts/source-sans-pro/tree/release/OTF
• fonts/source-serif-pro/* : See https://fonts.google.com/specimen/Source+Serif+Pro or https://github.com/adobe-fonts/source-serif-pro/tree/release/OTF
• fonts/eb-garamond/* : See https://bitbucket.org/georgd/eb-garamond/src/master
• fonts/pacifico/* : See https://fonts.google.com/download?family=Pacifico

On Linux, with make installed, the fonts can be set up with the following command:

make -C fonts

The fonts can be made discoverable by matplotlib (through fontconfig) by creating the following in $HOME/.config/fontconfig/fonts.conf (see here):

<?xml version="1.0"?>
<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
<fontconfig>
<dir>/path/to/cheatsheets/fonts/</dir>
...
</fontconfig>

2. You need to generate all the figures:

$ cd scripts
$ for script in *.py; do python $script; done
$ cd ..

3. Compile the sheet

$ xelatex cheatsheets.tex
$ xelatex cheatsheets.tex

cheatsheets's Issues

Missing ) in np.random.uniform

Thanks for the great cheatsheets!

I think there is a missing ) in

Z = np.random.uniform(0, 1, (8, 8)

under Matplotlib for Beginners/Choose (both imshow and contourf).

Replace `step` with `stairs`?

See the discussion in matplotlib/matplotlib#22277

No pcolormesh referenced

Folks should be usingpcolormesh in general rather than imshow - imshow really only works for equally spaced pixels, and there is lots of data that is not laid out that way. However the cheatsheet doesn't mention pcolormesh.

Variable names should generally be lowercase

In many places we have uppercase variables, e.g.

X = np.linspace(0, 2 * np.pi, 100)
Y = np.sin(X)

and likewise then in the plot functions.

This is against PEP-8 naming conventions and IMHO sets a bad example. At least 1D arrays should be lowercase. One may debate on 2D arrays, which have a history of using uppercase letters in the Matplotlib docs.

Image for "Basic plots - plot" should hint at markers

Currently the image is just sine line, which could trick users in thinking that plot is for lines and scatter is for markers.

I propose to additionally show another set of values with markers, e.g. something like:

markevery

handout-intermediate.pdf, "Lines & markers"
Given X of length 1000, the plot with markevery=25 should mark 1000/25 = 40 points, but there are 20 of them on the figure, so it seems a bug to me. (If so, it matters also to the next example, in "Scales & projections").

Fix latex code

Latex code is a bit messy, it would need some cleanup

A better answer about "How I rotate tick labels"

The cheatsheets comment the method "ax.set_[xy]ticks(rotation=90)" to rotate ticks. However, it can cause confusion because the parameter "takes effect only if you pass labels"(official tutorial of ax.set_[xy]ticks(). So this is a incomplete form and not recommended.
As the official tutorial says, I suggest ax.tick_params(axis='x', rotation=90).

Simplifying code

Most, if not all, np., plt., ax., fig., and mpl. can be safely removed if importing * directly in the main name namespace. Currently there are so many useless repetitions, IMHO.

from matplotpib.pylab import *

would make things much cleaner and simple.

Style the website

Website: https://matplotlib.org/cheatsheets/

The website is currently contains only plain HTML.
It would be better if we can make it look more like the official site: https://matplotlib.org/stable/index.html

Font issue

I can compile the tex document with the font specs commented out.
However, with the fonts defined as in the document, e.g

\setmainfont{Source Serif Pro}[
  Path = fonts/source-serif-pro/SourceSerifPro-, Extension = .otf,
  UprightFont= Light,
  ItalicFont = LightIt,
  BoldFont   = Regular,
  BoldItalicFont = It]

it is looking for fonts that do not exist. E.g.
fonts/source-serif-pro/SourceSerifPro-Light.otf is not part of any of the linked google font sets.

Update Beginner Cheatsheet: fig.show → plt.show

In the beginner handout, there is a brief section on how to display a given plot

https://github.com/matplotlib/cheatsheets/blob/05fbbb90103b1c68f4922096b1e7290b7dac1a17/handout-beginner.tex#L90-L95C9

I wanted to start a small discussion on whether the call to fig.show() should be replaced by plt.show.

Since this is a beginner cheatsheet, the management of a blocking GUI loop is slightly out-of-scope. fig.show does not manage this loop (whereas plt.show does), new users may wonder why their plot disappears as quickly as it appears when using fig.show.

'right' and 'top' in plt.subplots_adjust() are not directly padding size

Very thankful of the cheatsheet, but I think there's a little problem.
In 'Axes adjustments' of the second cheatsheet, 'right' directly refers to the padding between subplots' right edge and figure's right edge. However it actually refers to the relative distance between subplots' right edge and figure's left edge. The same also applies to 'top'.

Legend placement error

Legend placement numbers in the cheatsheet are inconsistent with matplotlib documentation. (e.g. for lower-right, loc=3 in the cheatsheet instead of loc=4).
https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.legend.html

`tight_layout` versus `constrained_layout`

cheatsheets/handout-tips.tex

Line 215 in 329e0ba

Once your figure is finished, you can call {\ttfamily tight\_layout()}

calls out tight_layout in particular. Is there a reason to not also call out constrained_layout? In general constrained_layout is more "automatic".

Make cheatsheet background non-transparent?

The handout thumbnails don't look very good on github's dark theme:

Add a top-level index.html

So that we have a nice target to point browsers to with links to the pdf and back to github.

This can probably be a static html page (not generated from sphinx).

Brazilian Portuguese translation

hello, there!
the PyLadies Brasil group is thinking about starting a Brazilian Portuguese translation of the cheatsheets and I wonder if there are specific guidelines for translations, i.e. where to save the translated files and so on, or if you are aware of any pt-br translation of this project that is ongoing.

if the answer is no for both questions, can we fork this repo and start? 😃

Automatic building of the PDF and the PNG files

It would be nice to have an automatic building of the PDF and PNG for each commit but I don't know if it is possible with e.g. Travis. It would require a partial TeXLive installation and I'm not sure how to setup.

Also, some have been manually post-processed with pdfcrop (to remove white margin) and this should be automatized.

cols, rows should be rows, cols

Under Subplots layout, the first argument is nrows, second is ncols

Update for 3.6.0

This requires updating the Sphinx theme (e.g. matplotlib/mpl-third-party#134), but also auditing the plots to make sure they did not break or could be using newer stuff.

The cheatsheets website should include a link to the GitHub page

Otherwise, the GibHub page is quite undiscoverable in case somebody wants to contribute.

A Suggestion about CI/CD.

Perhaps, we can use the GitHub Action or other CI/CD tools to automatically compile and release these beautiful and useful files.

Spanish translation

Hello, congratulations on the work everyone has made so far.
I'm looking forwards to translate the cheatsheets to Spanish.

I've been told to open an issue to gather those who want to help, so here it is.

Matplotlib Deprecation Warning in /scripts/scales.py

As we execute /scripts/scales.py, we have the following warning :

scales.py:27: MatplotlibDeprecationWarning: The 'basex' parameter of __init__() has been renamed 'base' since Matplotlib 3.3; support for the old name will be dropped two minor releases later.
  ax.set_xscale("log", basex=10)
scales.py:39: MatplotlibDeprecationWarning: The 'basex' parameter of __init__() has been renamed 'base' since Matplotlib 3.3; support for the old name will be dropped two minor releases later.
  ax.set_xscale("symlog", basex=10, linthreshx=1)
scales.py:39: MatplotlibDeprecationWarning: The 'linthreshx' parameter of __init__() has been renamed 'linthresh' since Matplotlib 3.3; support for the old name will be dropped two minor releases later.
  ax.set_xscale("symlog", basex=10, linthreshx=1)

Has GitHub Pages been activated for this repo?

The cheatsheets that are automatically built should be available here, but I am getting a 404. If GH Pages has not been activated for the repo, that could be the cause. There may also be a problem with the project pages and having a custom domain for the organization pages, as discussed in isaacs/github#547.

Missing name style

Thanks for the great cheatsheets!

The first page in Matplotlib cheatsheets, chart "style name" missing in "Style" - plot (2,1), that looks like "ggplot"

Missing closing parenthesis in beginners handout

I think some closing parenthesis are missing in the beginners handout:
https://raw.githubusercontent.com/matplotlib/cheatsheets/master/handout-beginner.png

When np.random.uniform() is used with the size argument as a tuple, the last closing parenthesis is missing.

Here:

and here:

Cursive font is not cursive

The line that should demonstrate a cursive font is printed with a regular sans fon (cheatsheet page 2 bottom left):

prune the number of image interpolations and replace with new material

Based on #80 (comment), there may be an opportunity to reduce the number of image interpolation methods that are demonstrated, and make room for new material.

In the handout for beginners, plt.subplots(2, 1) should be used.

I'm using matplotlib-3.6.0.

When I use plt.subplots((2, 1)), there is an error: ValueError: Number of rows must be a positive integer, not (2, 1)

We need a license

BSD ?

broken links on cheatsheets site

There are a couple of broken links in the navigation bar for the cheatsheets site https://matplotlib.org/cheatsheets, specifically the "Plot types" and "Release notes" links. Once current matplotlib dev version is released and the docs are published to https://matplotlib.org/stable, this would not be an issue; still, probably should be addressed.

Also, the search bar is a bit unnecessary, as it does not search the main site. Can the sidebar from the main site be integrated into the cheatsheets site. Or, maybe it is better to build the cheatsheets site in the matplotlib docs.

handout-beginner.pdf

Hello,

thank you for this great summary. In handout-beginner.pdf

Y1, Y1 = np.sin(X), np.cos(X)

should be

Y1, Y2 = np.sin(X), np.cos(X)

bytelinker

add candlestick cheat .

the most important thing on charting is finansal charts that matplotlib didnt work around it.
yes it have some libs and things but not sharply have do something for that.
please add cheatsheat for finansal and candlistic charts to help us thanks

Better to share vector version such as compiled PDF instead of jpg image

To save time for beginners but still give chance to old hands

Dash tuple is missing closing parenthesis

Thanks for publishing this cheatsheet. I learned about several features while reading through it.
In the Lines section, the example of using a dash tuple for linestyle is missing the closing parenthesis. I tried running your linestyles.py script, and it shows the closing parenthesis, so I guess it gets cropped when you assemble the whole thing in LaTeX.

Another minor suggestion: remove the /matplotlib from the end of the Stack Overflow help item, because it's not a valid URL. The PDF version of the cheatsheet has the full /questions/tagged/matplotlib in the link, but the PNG version might confuse people who try to type it in as listed.

Clarify when to use None / np.nan to segement data

see matplotlib/matplotlib#23714

These is some user confusion about when to use ax.plot with np.nan separated data and when something like LineCollection is a better choice. In particular, using the segmented data approach the line will be drawn as a single path which will not show self-intersection for alpha < 1.

Branch protection and merge rules?

Hey folks, do we want to enable branch protection? and do we want a certain # of reviews to merge stuff? attn: @rougier

Legend placement

Thank you for the awesome work!

Am I miss something? ... to me the strings for A-F are incorrect. The bbox_to_anchor seems fine though.

Cmap 'Blues' instead of 'Oranges'

There might be a typo in the tips & tricks handout where a Blues color map is called although an Oranges one is displayed:

For consistency, maybe we could use

cmap = plt.get_cmap("Oranges")

instead of

cmap = plt.get_cmap("Blues")

Typo in handouts tips

In the last paragraph we read documenation instead of documentation...

cheatsheets-2.png shows interpolation different for none vs None

If you look at the imagine cheatsheets-2.png linked in the README.md:
https://github.com/matplotlib/cheatsheets/blob/master/cheatsheets-2.png

in the interpolations image grid, the "none" version actually is using some sort of interpolation. This makes it seem like "none" and None are different, but they should not be.

When I run the interpolations.py script locally
(https://github.com/matplotlib/cheatsheets/blob/master/scripts/interpolations.py) I get the following:

interpolations.pdf

which has the correct top row.

The image in the repo should be updated.

Error with legend locations?

Think there may be an error in your cheatsheet. The legend locations seem to be different than those in the matplotlib documentation:

https://matplotlib.org/3.2.1/api/_as_gen/matplotlib.pyplot.legend.html#:~:text=matplotlib.pyplot.legend%20%C2%B6%20%20%20%20Location%20String%20,%20%203%20%207%20more%20rows%20

Issue with build process

I am attempting to build this repo with gitbash in Windows 10 pro. I have cloned the repo to a folder /git/cheatsheets and created /git/cheatsheets/fonts with all the required sub-folders as described in the README. However, when running the python scripts in the scripts directory I am getting a failure of

"findfont: Font family ['Source Code Pro'] not found. Falling back to DejaVu Sans."

I am guessing I have put the fonts folder in the wrong location and/or finding the fonts on the Windows file system is not working as expected.

Any help would be appreciated.

tight_layout is a method of plt, not ax

cheatsheets/cheatsheets.tex

Line 721 in af15065

ax.\textbf{tight\_layout}()\\

This should become plt.\textbf{tight\_layout}()\\

how to use

i dont know how to work, feel terrible

Cartopy version

@jimustafa what was the reason for pinning cartopy to 0.19.0.post1 in dbf92ab? It seems to work with any (newer) version too.

Is Delicious font necessary?

Its license is not a common Free one, though that doesn't preclude it from being Free. However, before getting into those weeds, is it really needed? I don't see 'delicious' mentioned anywhere but the readme.

Consistent capitalization of section titles

Capitalization of section titles switches between all words starting with capital letters and only the first word starting with a capital letter; e.g.

• Quick start
• Ten Simple Rules
• Ticks & labels
• Scales & Projections

Choose one way and stick with it throughout.

Note: The Matplotlib docs only capitalize the first word of a section. I don't care if the cheat sheets use the same rule, as long as they are consistent within themselves.