This repo is three things in one: 1- A set of instructions and design motifs AI agents can incorporate into their understanding of what makes a good Matplotlib graph through SKILLs and a CLAUDE.md. 2- An online 'blog' tutorial for a human audience to learn the same skills (to remind me how to do specific things or teach an intern/coworker the same idea). 3- An included mplstyle file that contains my opinionated set of sane default configurations for matplotlib: minerva.mplstyle.

  This repo is three things in one:
  1- A set of instructions and design motifs AI agents can incorporate into their understanding of what makes a good Matplotlib graph through SKILLs and a CLAUDE.md.
  2- An online 'blog' tutorial for a human audience to learn the same skills (to remind me how to do specific things or teach an intern/coworker the same idea).
  3- An included mplstyle file that contains my opinionated set of sane default configurations for matplotlib: minerva.mplstyle.

My biggest inspiration was Tufte's The Visual Display of Quantitative Information which my boss gifted me a couple years ago, data-to-viz(dot)com and the python-graph-gallery were the big references for "which chart for which data", and to be fair those are by the same author as a paid course I couldn't justify paying for to myself called matplotlib-journey(dot)com. My is to try to get half the way to the amazing plots they show you but free, open-source so I can share it with friends

I'd genuinely be interested just to use this post as a way to survey the HN community on what things levelled-up your own Matplotlib graphing or what finally made general science/quantitative data communication click.