Figures for Papers is a Python collection of scripts for creating publication-ready scientific figures.
The repository addresses the challenge of producing high-quality, visually consistent figures for academic papers submitted to top-tier venues. It provides reusable Python scripts organized by figure type—including bar plots for quantitative comparison and composition breakdown, radar plots, line plots, 3D visualizations, concept diagrams, and trend plots. Each script is designed to generate figures meeting the standards of major conferences and journals in machine learning and computer science.
Researchers preparing submissions to venues like NeurIPS, ICML, ECCV, or Nature Machine Intelligence should consider this collection if they want to accelerate figure creation and maintain visual consistency across papers. The tool is particularly suited for those comfortable writing Python and seeking to avoid repetitive plotting work. The repository includes both end-to-end Python implementations and acknowledgment of figures created with partial Python assistance, giving users flexibility in how they integrate the scripts into their workflow. The project also exposes its figure-making approach as a reusable skill that can be referenced directly in AI coding agents without installation, allowing developers to prompt AI assistants to generate plotting code following the documented design principles.
Development activity shows consistent engagement with the repository's educational mission. The maintainer has organized scripts by project and figure type, making it straightforward for others to locate and adapt relevant examples. Documentation includes design theory references and skill definitions formatted for AI agent integration, indicating intentional effort to make the work accessible beyond direct code reuse. The inclusion of a skill folder hierarchy and explicit guidance for using the collection with AI coding agents demonstrates responsiveness to how modern development workflows operate.
