clio-clustering

Builds an end-to-end pipeline to cluster text data with HCI and visualize it.

Updated Dec 30, 2025

Search Tags:#visualization #nlp #embedding #clustering #data-sourcing #text-analysis

Ready to run

npx skills add https://github.com/josh-cooper/.claude --skill clio-clustering

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: clio-clustering
Source: https://github.com/josh-cooper/.claude/tree/main/skills/clio-clustering
Command: npx skills add https://github.com/josh-cooper/.claude --skill clio-clustering

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Builds an end-to-end clustering and visualization pipeline for text data, enabling pattern discovery and exploratory analysis across sources like GitHub issues, Slack messages, tickets, code reviews, and forum posts.

Core Features & Use Cases

Data sourcing, scraping, embedding, and hierarchical clustering to group similar items.
LLM-powered cluster labeling and descriptive summaries for quick interpretation.
Interactive visualization to explore clusters, filter results, and drill down to individual items.

Quick Start

Follow the phases to identify data sources, verify access with tests, fetch data, generate embeddings, perform hierarchical clustering, label clusters with an LLM, and launch the interactive visualization.

Frequently Asked Questions about clio-clustering

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

FAQPage Schema

How do I cluster and visualize text data from GitHub issues and Slack messages?▼

To cluster text data from sources like GitHub issues and Slack messages, this pipeline scrapes your sources, generates OpenAI embeddings, and applies HDBSCAN hierarchical clustering. It then produces an interactive visualization to explore the resulting clusters and drill down into individual items.

What is the best way to automatically label clusters of unstructured text?▼

Using an LLM to label clusters of unstructured text automatically generates descriptive summaries for quick interpretation. This approach pairs hierarchical clustering results with LLM-based labeling to assign meaningful titles and descriptions to each discovered group.

Do I need an OpenAI API key to generate text embeddings for clustering?▼

Yes, you need an OpenAI API key set as an environment variable to generate text embeddings. The semantic clustering pipeline relies on the OPENAI_API_KEY environment variable to process text data, alongside provider-specific tokens for data sourcing.

Can I use HDBSCAN hierarchical clustering for exploratory data analysis?▼

Yes, HDBSCAN hierarchical clustering is well-suited for exploratory data analysis of text embeddings. It groups similar items into meaningful clusters without requiring a predefined number of clusters, making it ideal for discovering unknown patterns in text data.

Does this semantic clustering pipeline support data sourcing from forums and tickets?▼

Yes, the pipeline supports data sourcing and scraping from forums, tickets, code reviews, and Slack messages. It fetches text data directly from these sources before generating embeddings and performing hierarchical clustering for exploratory analysis.

Why does my text clustering workflow need a SQLite backend?▼

A SQLite backend manages the workflow state and fetched text data during the clustering pipeline. It supports the end-to-end process by storing sourced data, embeddings, and hierarchical clustering results before rendering them in the interactive visualization frontend.