← ClaudeAtlas

researchzy-figureslisted

Make publication figures from data: profile the data, choose a chart it can support, draw it with matplotlib at the venue's column width, export PDF and PNG with a record, and check the file's size, fonts and colours. Use for a new figure, or to check a figure file before submission.
Researchzy/Researchzy-academic-skills · ★ 0 · Data & Documents · score 72
Install: claude install-skill Researchzy/Researchzy-academic-skills
# Researchzy Figures This skill turns a data file into a figure for a paper and then checks the exported file, not the code that drew it. The order is fixed: agree what the figure must show, profile the data, choose a chart the data can support, draw it from a script at the printed size, export it with a record of what it was made from, run the checker, and look at the result. It draws with Python (matplotlib, optionally seaborn). The checker reads any exported PDF, PNG or TIFF, so it also checks figures made in R or a drawing program. The skill does not use image-generation models, and the only diagrams it draws are flow charts drawn by code. It does not run statistical tests or decide what an error bar should be. Resolve `SKILL_DIR` to the directory containing this `SKILL.md`. The scripts need Python 3.10 or newer. Drawing needs matplotlib 3.6 or newer; the profiler and the checker need nothing else, and the checker compares colours only when Pillow is installed. If matplotlib is missing, say so and ask before installing anything. No account is needed; the Researchzy tools are not used. ## Inputs - A data file with a header row: CSV, TSV or Excel (Excel needs pandas). One row per observation is best. A table of summaries is accepted, but then no distribution can be drawn. - What the figure should show and where it will be published. - Or a figure file that already exists: go to "Checking a figure made elsewhere". ## Workflow Everything goes in `figures/` in the user'