figures4papers

Why we included this project

Graduate students and researchers who have spent an evening fighting matplotlib to get a figure looking conference ready will recognize what this is: a working collection of scripts behind the actual charts in papers accepted at Nature Machine Intelligence, ICML, NeurIPS, and ECCV. Rather than reverse-engineering publication style from journal guidelines, you can grab a proven bar comparison, radar, line, trend, or 3D sphere script and adapt it to your own data. The repo also bundles a structured figure-making skill with style notes and API references that an AI coding agent such as Cursor or Claude Code can follow, so it works as both a manual reference and a promptable guide inside an editor. One honest caveat: this is a personal toolkit built around one author's papers, not a packaged library, so expect to customize rather than drop it in.