data-viz-plots_skill

Overview

This skill creates publication-quality scientific visualizations using matplotlib and seaborn, executed locally in your environment. It supports common plot types for multi-omics workflows—scatter, heatmap, violin/box, UMAP/tSNE, dot plots, volcano plots and multi-panel figures—and works with any LLM provider. The emphasis is on reproducible, high-resolution figures ready for papers, posters, and presentations.

How this skill works

The skill provides ready-to-run plotting patterns and snippets that accept pandas DataFrames, numpy arrays, or AnnData objects and render figures with consistent styling. It sets sensible defaults for DPI, fonts, palettes and layout, offers utilities for multi-panel layouts and custom color mapping, and includes export commands to save high-resolution PNG/SVG files. Troubleshooting tips and density/aggregation approaches help with large, noisy datasets.

When to use it

• Preparing figures for manuscripts, posters, or slide decks
• Exploratory data analysis (QC metrics, clustering, marker genes)
• Visualizing single-cell, bulk, or spatial RNA-seq results (UMAP, heatmaps, dot plots)
• Generating multi-panel composite figures combining different plot types
• Exporting high-resolution images for publication or downstream editing

Best practices

• Choose figure size and DPI for the target medium (papers: 6–8 in wide, save at 300 DPI)
• Use colorblind-friendly palettes (viridis, Set2, tab10) and define palettes once for consistency
• Apply alpha and smaller point sizes to reduce overplotting in dense scatter plots
• Call plt.tight_layout() and use bbox_inches='tight' when saving to avoid clipped labels
• Use vector formats (SVG) for editable graphics and PNG for raster needs
• Close figures (plt.close()) when generating many plots to free memory

Example use cases

• QC dashboard: histograms, scatter UMI vs genes, and mitochondrial violin plots for single-cell QC
• UMAP cluster visualization: colored cluster scatter plots with external legend and reproducible palette
• Gene expression heatmap and dot plot for marker gene panels across clusters
• Differential expression volcano plot with thresholds and category coloring for DEG reports
• Multi-panel figure combining UMAP, violin, and heatmap for a single-figure result in a manuscript