generate-ingredient-umap

Generate interactive UMAP visualizations of CultureMech CHEBI ingredient embeddings.

10|3|Updated Jan 28, 2026
One-click install
npx skills add https://github.com/CultureBotAI/CultureMech --skill generate-ingredient-umap
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: generate-ingredient-umap
Source: https://github.com/CultureBotAI/CultureMech/tree/main/.claude/skills/generate-ingredient-umap
Command: npx skills add https://github.com/CultureBotAI/CultureMech --skill generate-ingredient-umap

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It turns CultureMech CHEBI ingredient embeddings into an explorable map so you can quickly understand ingredient relationships, frequency, and coverage gaps without manually sifting through raw graphs or embedding files.

Core Features & Use Cases

  • Interactive UMAP visualization: Generates a browser-ready HTML scatter plot (app/ingredient_umap.html) for 2D exploration of CHEBI ingredients.
  • Tiered, frequency-aware insights: Colors points by occurrence tier and sizes them by log-scale frequency to highlight the most important ingredients.
  • Actionable tooltips and search: Shows ingredient name, CHEBI ID, CAS-RN (when available), occurrence count, and example media, with filtering by ingredient name or CHEBI ID.

Quick Start

Run the full visualization generation with the local embeddings by executing: just gen-ingredient-umap

Frequently Asked Questions about generate-ingredient-umap

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I visualize CHEBI ingredient embeddings in an interactive HTML plot?

You can visualize CHEBI ingredient embeddings by running a script that reduces KG-Microbe DeepWalk vectors into a 2D UMAP scatter plot. This generates a browser-ready interactive HTML file for exploratory analysis.

What is UMAP used for in ingredient discovery and coverage auditing?

UMAP projects high-dimensional ingredient embeddings into a 2D space, allowing you to visually explore ingredient relationships, identify frequency-driven priorities, and detect coverage gaps across culture media.

Do I need local embedding TSV files to generate the UMAP visualization?

Yes, the visualization requires local embedding TSV inputs. The script extracts CHEBI-only ingredients and uses cached embedding handling to ensure deterministic HTML output and optimal performance.

How does the interactive UMAP plot handle ingredient frequency and metadata?

The plot colors points by occurrence tier and sizes them by log-scale frequency. Tooltips display ingredient name, CHEBI ID, CAS-RN, occurrence count, and example media, with text filtering capabilities.

Can I use D3.js to interactively filter CHEBI ingredients by name or ID in the browser?

Yes, the generated interactive HTML visualization includes search functionality that allows you to filter ingredients by name or CHEBI ID directly within the browser-based D3.js plot.

What are the limitations of using offline UMAP visualization for ingredient analysis?

This approach is limited to offline visualization workflows using static embedding inputs. It does not support real-time data updates or dynamic embedding generation from live data streams.