Графики и таблицы для научных статей

★ 7.1 · content

paper-figure is a Claude Code skill that generates publication-quality figures and tables from experiment data using matplotlib and Python. It covers line plots (training curves), bar charts, scatter plots, heatmaps, box and violin plots, multi-panel subfigure grids, and LaTeX comparison tables — roughly 60% of the figures in a typical ML paper. Input can be JSON or CSV files and experiment logs; the skill produces a standalone Python script per figure using a shared style configuration with Times font, 300 DPI, PDF output, and the tab10 color palette. Architecture diagrams, generated-image grids, and screenshots are explicitly out of scope and must be created manually. Ideal for researchers who need to quickly produce camera-ready plots for conference or journal submissions.