seaborn

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• plot this data with seaborn
• violin plot
• pair plot
• seaborn heatmap
• distribution plot

What it does

seaborn is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a seaborn expert for statistical visualization directly from pandas DataFrames — covering distribution plots (histplot, kdeplot, ecdfplot), categorical plots (boxplot, violinplot, stripplot, swarmplot), relational plots (scatterplot, lineplot), matrix plots (heatmap, clustermap), and the figure-level interfaces (FacetGrid, PairGrid).

A session produces clean, legible Python code that goes from a DataFrame to a publication-worthy exploratory figure with appropriate statistical annotations and sensible aesthetic defaults.

When to use it

Reach for it when:

• You want to explore distributions, relationships, or group differences in tabular data quickly
• You’re generating figures for papers or reports where seaborn’s built-in statistical aggregation (confidence intervals, regression lines) saves you from writing the aggregation code yourself
• You need pair plots or faceted grids to compare multiple variables simultaneously

When not to reach for it:

• Fine-grained custom styling at the Artist level — use matplotlib directly
• Publication-quality multi-panel figures with journal-specific requirements — use scientific-visualization
• Interactive web plots — consider plotly

Install

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

Trigger phrases: “plot this data with seaborn”, “violin plot”, “pair plot”, “seaborn heatmap”.

What a session looks like

A typical session has three phases:

1. Data and plot specification. Describe your DataFrame structure and the comparison or distribution you want to visualize. Claude identifies the most appropriate seaborn plot type and figure-level interface.
2. Code generation. Claude writes seaborn code with explicit theme setting (sns.set_theme()), palette selection, and axes-level plot calls that can be embedded in a matplotlib figure for multi-panel composition.
3. Statistical annotation. If comparisons between groups are needed, Claude adds significance annotations using statannotations or manual matplotlib annotation on top of the seaborn base plot.

Receipts

Where it works well:

• Pair plots for exploratory analysis of multi-variable datasets — seaborn’s PairGrid handles the faceting and diagonal plots reliably with a single function call
• Violin + strip plot combinations for showing both distribution shape and individual data points — a common pattern in biology that seaborn makes straightforward

Where it backfires:

• Very large datasets where seaborn’s row-by-row rendering becomes slow — for datasets over ~100k points, switch to hex bins or use datashader
• Complex multi-panel layouts where panels need to share axes or have different plot types — the matplotlib subplot interface is cleaner for these

Pattern that works: use seaborn for the exploratory phase to understand your data, then port the best plots to scientific-visualization for the final publication figure with journal-specific styling applied.

Source and attribution

Originally authored by K-Dense Inc.. The canonical SKILL.md lives in the seaborn 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.