ml-cluster - Find natural groups in data and label them
Find natural groups in a dataset using unsupervised clustering, then label each group with a plain-language description.
Tags
Updated: 2026-09-19Capabilities
What this skill does
- Check environment and load dataset
- Generate feature or text representations
- Reduce representation dimensionality using UMAP
- Cluster data using unsupervised algorithms
- Generate plain language cluster labels
- Present formatted clustering result summary
- Save and resume execution checkpoints
Inputs
- Data file or text directory path
- Optional natural language clustering objective
Outputs
- Formatted clustering result presentation
- Checkpoint files in checkpoint directory
- Representations in shared embedding cache
Requirements
- Python 3 environment with uv
- pandas package
- scikit-learn package
- LLM access for label generation
- Optional sentence-transformers package
- Optional umap-learn package
- Optional hdbscan package