plotly-interactive-plots

What it does

The skill trains an AI to generate interactive Plotly visualizations for scientific data, using two APIs: plotly.express (px) for high-level, one-line DataFrame plots and plotly.graph_objects (go) for fine-grained control over traces. It covers more than 40 chart types with features like hover, zoom, pan, and animation, and explains exporting figures as HTML or static images via kaleido. It illustrates use cases such as volcano plots with gene hover, dose-response dashboards, heatmaps, and 3D molecular views, and notes when to use seaborn or matplotlib for specific needs.

How it works

• Uses plotly.express (px) for quick, DataFrame-to-plot workflows (scatter, line, violin, histogram, heatmap via imshow) and automatically generates hover tooltips.
• Uses plotly.graph_objects (go) for full trace-level control, enabling custom classifications (significant vs non-significant), multiple traces, and threshold lines.
• Demonstrates core modules:
  • Module 1: px Scatter and Line — Relational Plots, with examples of dose-response scatter and time-course expression plots.
  • Module 2: px Statistical Plots — Distributions and Categories (box, violin, histogram, strip).
  • Module 3: px Heatmap and Matrix — Gene expression heatmaps and correlation matrices via imshow.
  • Module 4: go Graph Objects — Full Trace Control (volcano plot with classifications, bar chart with error bars).
  • Module 5: 3D and Specialized Charts — 3D PCA scatter and parallel coordinates.
  • Module 6: Subplots and Export — make_subplots usage and exporting figures with write_html and write_image (kaleido).
• Includes concrete code blocks with exact function calls and parameters, showing how to create, customize, and export figures.

When to use it

• When you need hover tooltips with gene names, p-values, or metadata without cluttering a static figure.
• When building multi-panel interactive dashboards for dose-response curves, cohorts, or multi-condition comparisons.
• When sharing figures as self-contained HTML files for browser exploration.
• When creating 3D or higher-dimensional visualizations (3D scatter, surface, parallel coordinates).
• When you require animation frames or specific publication-ready tweaks, and you know to fall back to seaborn or matplotlib for certain statistical or print-quality needs.

What it can touch

• Tools: "plotly.express as px", "plotly.graph_objects as go", "plotly.subplots.make_subplots".
• Code examples manipulate: figures, traces, hover templates, axes, layout, and export methods (write_html, write_image).

Caveats

• Requires external packages for static image export: kaleido (installation shown).
• Uses specific data structures (DataFrames, arrays) and sample datasets; real usage should supply appropriate data.
• The skill does not guarantee any particular visualization outcome or performance across environments.