Introducing Wolfe, A More Modern, Edgy, Visual Style for Python's Matplotlib


     Having began my exploration of technical programming with Mathematica, I was introduced to probably the best default plot visuals of most CAS programs and certainly python's matplotlib. However as my experience and interest in Python grew, I wanted to create a style sheet that I could revert to that would render beautiful plots by default without any in-line coding required, especially plots that I could print. Hence, grew the Wolfe style sheet project. Its still a work in progress, to refine it for the variety of plots one may ever create through matplotlib, but overall, its use-ready now. The project can be found on my GitHub Repo page.

     The main points I found to be problematic using the included style sheets, especially the default, was that I found them visually crude. In particular, I disliked font choices, line and marker widths as well as the color schemes and plot sizes. Another, larger issue was that none of the style sheets were adequate to print large plots by default. To tweak all of these items in-line to my liking I may as well just write an entire style sheet. To give a visual comparison of what I mean:

Default Seaborn Wolfe

    

     A better comparison can be seen on the repo's wiki page or you can download the image files themselves there as well. There are actually two style sheets. The standard style sheet generates a moderately sized image (681 x 545 px) meant for working on electronic media presentations, like the web or powerpoint/impress. However the way the DPI works in the code, its unsuitable for printing all but the smallest of images. Therefore, a high-res version is also included. The high resolution version renders a gigantic photo (2549 × 2044 px) that prints beautifully at 10" x 8" size. The standard Wolfe style sheet is meant to be used for in-line coding as well, such as displaying in a Conda notebook. The large size of the high-res version means that it usually has problems being displayed in this fashion and should only be saved as a figure in the python code. All of these technical requirements (there aren't many) are explained on the repo's wiki page.

     As far as the visual goes, its based upon the included BMH style. However I dislike the line widths of the plots themselves - I think they're too heavy. As previously stated I also dislike the font choices as well as the tick-mark display and color choices. In any regard, everyone can form their own opinions but I think tweaking any of the other styles to be print worthy would take too much time, and would have to be done every time. This way, one just imports the style sheet once. Tweak the colors from there if you wish. Leave a note on the github and tell me what you think as well as give me any suggestions!

© 2016 Zachary G Wolfe -- Remember to turn your brain off for a reboot sometimes...