Making Plots Pretty
PrefacePermalink to “Preface”
Back in 2014, I gave this talk to my robotics lab on how to make good looking plots / figures. I’m posting it now in 2021. It’s fun to remember how far back my obsession with making figures goes.
If you’re interested in using this material, the Keynote file is available in a GitHub repository:
A 2014 lab talk I gave on how to make good looking plots and figures.
– Max from 2021
SlidesPermalink to “Slides”
AcknowledgmentsPermalink to “Acknowledgments”
Thanks to the following folks whose resources I used either as examples of pretty things, or to make my own (somewhat) pretty things.
- Bilge Mutlu
- Ola Botvinnik (esp PrettyPlotLib)
- Cynthia Brewer (esp Colorbrewer)
- John Hunter & The Matplotlib Team
- Loren Shure (esp this blog post)
- Matt Davis (esp brewer2mpl)
PostscriptPermalink to “Postscript”
(Max from 2021 again here)
I couldn’t get other plot configurations or libraries to work, so I packaged my own set of plot modifications into a function and released it.
My own attempt at a little library to make bar plots prettier
Please don’t use it. The standard configurations and libraries are assuredly the way to go. Even my guide for setting up this library (changing
PYTHONPATH) is bad. I think a great modern alternative to this genre of libraries is Seaborn.
At the same time, I thought it would be fun to include here because it shows how many axes of growth there are in developing even little tools like this:
- configuration (e.g., why I couldn’t get the existing approaches to work)
- breadth (e.g., I only supported bar plots)
- aesthetics (e.g., I don’t think graying out many axis markers is actually helpful)
- packaging (e.g., my setup instructions are all wrong)