scientific-visualization

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• make a publication figure
• format for Nature style
• multi-panel figure
• journal-ready plot
• add significance annotations

What it does

scientific-visualization is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a publication figure specialist that orchestrates matplotlib, seaborn, and plotly to produce multi-panel figures meeting specific journal requirements — Nature, Science, Cell, and others — with correct font sizes, colorblind-safe palettes, significance bars, error bars, and figure dimensions.

A session produces Python code that generates a figure to spec: the code handles subplot layout, styling, annotation, and export at the correct DPI and file format for the target journal’s submission system.

When to use it

Reach for it when:

• You have data ready and need a figure that will pass journal technical requirements without revision requests
• You’re building a multi-panel figure (A/B/C/D panels) and need consistent styling across panels
• You need colorblind-accessible color schemes with significance annotations applied correctly

When not to reach for it:

• Quick exploratory data visualization — use seaborn or matplotlib directly
• Structural diagrams and flowcharts — use scientific-schematics

Install

Copy the SKILL.md from K-Dense AI’s scientific-visualization folder into .claude/skills/scientific-visualization/ in your project.

Trigger phrases: “make a publication figure”, “format for Nature style”, “multi-panel figure”, “journal-ready plot”.

What a session looks like

A typical session has three phases:

1. Journal and figure specification. Specify the target journal and figure type. Claude retrieves the journal’s figure guidelines (column width, DPI, font size, color mode) and proposes a panel layout.
2. Code generation. Claude writes matplotlib/seaborn code with publication styling: Helvetica or Arial fonts at correct point sizes, colorblind-safe palette, significance annotation for statistical comparisons, and tight layout for multi-panel figures.
3. Export and validation. The figure is saved at the correct DPI and format (TIFF/EPS/PDF), and the code includes a dimension check that warns if output size doesn’t match journal requirements.

Receipts

Where it works well:

• Cell-style violin plots with overlaid jitter and significance bars — the styling code is comprehensive and the output passes technical review without manual adjustment
• Colorblind-safe palette enforcement — Claude applies Wong or Okabe-Ito palettes consistently rather than defaulting to red/green

Where it backfires:

• Journals with proprietary style templates or EPS-specific font embedding requirements may need manual post-processing
• The skill generates code, not rendered figures; if your data doesn’t match the assumed format, you’ll need to adjust the data-loading step

Pattern that works: provide a small sample dataset (5–10 rows) alongside the figure request so Claude can write data-loading code that matches your actual structure rather than using placeholder arrays.

Source and attribution

Originally authored by K-Dense Inc.. The canonical SKILL.md lives in the scientific-visualization folder of their public scientific-agent-skills repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.