CorrelAid matplotlib style sheet and helper functions

This repository provides a matplotlib style sheet following the CorrelAid style guide and some plotting helper functions, possibly useful in the CorrelAid context, i.e. talks, blog posts, and such. It was created in the context of a blog entry.

Besides the default style, matplotlib comes with several built-in styles that we can use readily. To see a list of the available styles use:

import matplotlib.style as style
print(style.available)

['seaborn-deep', 'seaborn-muted', 'bmh', 'seaborn-white', 'dark_background','seaborn-notebook', 'seaborn-darkgrid', 'grayscale', 'seaborn-paper', 'seaborn-talk', 'seaborn-bright', 'classic', 'seaborn-colorblind', 'seaborn-ticks', 'ggplot', 'seaborn', '_classic_test', 'fivethirtyeight', 'seaborn-dark-palette', 'seaborn-dark', 'seaborn-whitegrid', 'seaborn-pastel', 'seaborn-poster']

For instance, with the 'ggplot' one can use a style that is similar to the default R ggplot style:

style.use('ggplot')

Here, we contribute a CorrelAid theme and helper functions. The package can be used in two different ways:

• Option 1: One can use the style file as described below on its own to style plots.
• Option 2: One can install this repository as Python package and use the style and the helper functions.

Install

Option 1 Using only on the style file

The stlye file should be placed in your user folder in the .matplotlib folder (should already exist), in a folder stylelib as follows:

~/.matplotlib/stylelib/correlaid.mplstyle

You may verify that it is detected by using the above style.available command.

Alternatively, you can simply put the file to your working directory and use the absolute path.

Option 2 Install using pip

To install the package simply run

pip install correlaidmatplotlib

This installs the package locally using pip and installs matplotlib, if not available, as well as numpy and seaborn that are only used for the example notebooks.

Option 3 Manually download and locally install the correlaidmatplotlib package

This is similar to option 1, and maybe useful if you want to add changes to the package. Then download or checkout this repository. Then in the top level that contains the setup.py file, run

pip install .

This installs the package locally using pip and installs matplotli, if not available, as well as numpy and seaborn that are only used for the example notebooks.

For all three options

The style uses the Google Roboto font. Make sure it is installed in your system. If there is a error message that the font cannot be found (may happen in Jupyer notebook), try deleting the matplotlib cache directory.

Use

Option 1 Using only the style file

The the style file can then be used as follows:

style.use('correlaid')

If not added to the user directory, it can also be placed in the working directory and linked:

style.use('./correlaid.mplstyle')

Then the following default color cycler is used for plotting:

import seaborn as sns
import matplotlib.pyplot as plt
sns.palplot(plt.rcParams['axes.prop_cycle'].by_key()['color'])

So for instance, a Seaborn plot in the default style looks as follows

import seaborn as sns
tips = sns.load_dataset("tips")
ax = sns.boxplot(x="day", y="total_bill", hue="smoker", data=tips)

It can be changed to the CorrelAid style to look as follows:

import matplotlib.style as style
style.use("correlaid")
ax = sns.boxplot(x="day", y="total_bill", hue="smoker", data=tips)

Options 2 and 3 Useing the package and the helper functions

When using the Python package, then one can simply import the package to apply the style, without using style.use('correlaid').

import correlaidmatplotlib

Note:: the import should be after importing matplotlib and seaborn. Otherwise and explicit style.use('correlaid') (see Option 1), should be inserted after all the imports and other style uses, to make sure the font sizes are correct.

While the default Seaborn plots are quite powerful and the style already improves some defaults, there is still room for improvement. With the CorrelAid package we provide helper functions to further style the plots consistently, i.e., we:

• add Title and subtitle,
• add a signature bar adding a Copyright and data source,
• take away the x and y axis labels, that are not necessary, when using the subtitle,
• move the legend to the right of the figure

fig, ax = plt.subplots()
sns.barplot(x="day", y="total_bill", hue="smoker", data=tips, ax=ax)
set_correlaid_style(fig, "Tipping Habits of Smokers", "Bill total, in US Dollar", source="tips.csv",
                   title_x_pos=0.03, subtitle_x_pos=0.03, copyright_x_pos=0.03,
                   title_y_pos=0.1, subtitle_y_pos=0.0, signature_fontsize="smaller")

For more control over the individual parts, see the notebook Option2 Install Package.ipynb.

Contribute

The style file is not yet complete and will only work for a subset of all the possible matplotlib and Seaborn plots. There may still be lines and parts of the plot that may not have been correctly styled yet. Feel free to add more parameters to the file via a pull request or raise an issue for plot types that appear not yet as expected. Of course any feedback on styles, or ideas for improvements are welcome.

More Information

The style is mostly based on this blog post:

• https://www.dataquest.io/blog/making-538-plots/