SpiffyPlots

A collection of opinionated matplotlib style sheets and plotting tools for publication-ready figures.

Requires Python 3.10 or newer and matplotlib 3.8 or newer.

• Free software: MIT license
• Documentation (likely not up to date): https://spiffyplots.readthedocs.io.

Installation

Install the latest release:

pip install spiffyplots

or install the latest commit directly from GitHub:

pip install git+https://github.com/JRBCH/spiffyplots.git

Using the style sheets

Import spiffyplots and load the default spiffy style to get more sensible default plotting parameters across the board.

What does it do?

Life is too short to look at ugly figures. This style fixed that problem for me, and over time I added a bunch of quality-of-life features to it, such as:

• Editable vector text by default. PDF and PostScript files embed TrueType fonts, and SVG retains <text> elements instead of converting glyphs to paths.
• Sensible default colors, fonts, and font sizes.
• Zero outer padding for better control over the final figure size.

Combining style sheets

The base spiffy style contains my own rather opinionated preferences for everyday use. You can combine it with other style sheets to match it to your use case.

import matplotlib.pyplot as plt
import spiffyplots

plt.style.use("spiffy")  # base style only
plt.style.use(["spiffy", "tol-muted"])  # plus a color scheme
plt.style.use(["spiffy", "nature"])  # plus a journal-specific style

[!NOTE] Import spiffyplots before calling plt.style.use. The import registers the style sheets.

What styles are included?

Why journal-specific style sheets?

Because journal guidelines are oddly specific and differ in some important details (e.g. whether a double-column figure is 175mm or 183mm wide), and I wanted a set-and-forget way of defining the journal style once, so I can focus on the actual data visualization. The API for this might change in future releases.

How to use Journal-specific styles:

Journal styles automatically apply journal guidelines. They set a default figure dimension (double column by default), correct font sizes for all plot elements, and export resolution. Some features, such as different figure size options, can instead be accessed through the spiffyplots.journals module. For example, spiffyplots.journals.nature.single_column(height=65, unit="mm") returns a figure size in inches that matches Nature's single-column figure guidelines with your specified height.

Multi-Panel figures

matplitlib's GridSpec is great, but I got real tired of formatting and re-formatting panel alignments. MultiPanel is my attempt of making this process easier.

import matplotlib.pyplot as plt
from spiffyplots import MultiPanel

plt.style.use("spiffy")

figure = MultiPanel(
    grid=[4, 2],  # 4 panels in the first row, 2 in the second
    labels=True,  # label panels (a, b, c, ...)
)

# Access each subplot axis by its label
figure.panels.a.hist(data)
figure.panels.b.scatter(x1, x2)
(...)

See examples/ex_multipanel.py for the full example and a comparison to the default matplotlib style. There are many other ways to define MultiPanel layouts, label placements, grid arrangements etc, see the example notebook for a short tutorial.

[!NOTE] Since I wrote this module, matplotlib has added subplot_mosaic, which solves a very similar problem. I may rework MultiPanel to wrap that instead of GridSpec one day, or stop using it all together, but for now I still prefer the spiffy interface.

Colors and colormaps

Style sheets only set the default cycle. For easy access to sensible colors, I use spiffyplots.colors for discrete colors and spiffyplots.cmap for colormaps. I really love Paul Tol's color schemes, so I heavily rely on them.

import spiffyplots as spiffy

scheme = spiffy.colors.tol_vibrant
color1 = scheme.red  # '#CC3311'
color2 = scheme.blue  # '#0077BB'

spiffy.colors.from_cmap(spiffy.cmap.sequential, 6)  # 6 equally spaced colors
spiffy.colors.shades(color1, 5)  # 5 shades of one base color

Schemes: tol_bright, tol_high_contrast, tol_vibrant, tol_muted, tol_medium_contrast, tol_light, tol_pale, tol_dark and okabe_ito. Each is a tuple you can index, slice and iterate, whose colors are also reachable by name. Each runs in its author's recommended order, then black, then grey.

Colormaps

Again, I wrapped Paul Tol's colormaps together with the default matplotlib colormaps in a convenient way for access & modification:

cmap = spiffy.cmap.sequential  # default sequential colormap (Paul Tol's iridescent)
cmap = spiffy.cmap.diverging  # default diverging colormap (Paul Tol's nightfall)
cmap = spiffy.cmap.iridescent_r  # a specific colormap, reversed
cmap = spiffy.cmap.viridis  # falls through to matplotlib

You can also create colormaps from one or more base colors:

cmap = spiffy.cmap.from_base(color1)  # white -> color -> darker color
cmap = spiffy.cmap.from_colors(spiffy.colors.tol_muted)

Perceptually uniform colormaps

cmcrameri and Colorcet do this exceptionally well, and spiffy.cmap can access their colormaps once either is installed. To install them together with spiffyplots:

pip install "spiffyplots[colormaps]"

You can then access them through spiffy.cmap as well:

import spiffyplots as spiffy

cmap = spiffy.cmap.batlow  # cmcrameri, registered as `cmc.batlow`
cmap = spiffy.cmap.kbc  # colorcet, registered as `cet_kbc`

Credits

• Most color schemes are Paul Tol's, redistributed under the 3-clause BSD license.

• Shoutout to cmcrameri and Colorcet for their beautiful perceptually uniform colormaps.