An interactive heatmap visualization tool for exploring tabular data with on-the-fly clustering, correlation networks, pathway analysis, and an AI-powered chatbot. Built for researchers and data scientists.
Live demo: https://clusterchirp.mssm.edu
The easiest way to run ClusterChirp locally is with Docker. No setup required.
docker pull ghcr.io/gumuslab/clusterchirp:latest
docker run -p 8080:80 ghcr.io/gumuslab/clusterchirp:latestThen open http://localhost:8080 in your browser.
The built-in AI chatbot lets you interact with your heatmap using natural language commands (e.g., "show top 20 most variant genes", "cluster rows", "sort by variance"). It requires an OpenAI API key.
docker run -p 8080:80 -e OPENAI_API_KEY=sk-your-key ghcr.io/gumuslab/clusterchirp:latestYou can get an API key at https://platform.openai.com/api-keys.
All other features (heatmap visualization, clustering, correlation networks, pathway analysis) work without an API key.
- Interactive Heatmap — Upload CSV, TSV, or XLSX files and visualize with real-time clustering
- AI Chatbot — Natural language commands to filter, sort, cluster, and explore your data
- Correlation Networks — Visualize gene correlation networks in 2D and 3D
- Pathway Analysis — Enrichment analysis with KEGG, Reactome, WikiPathway, and more
- Example Datasets — Preloaded proteomics, genomics, and immunogenomics datasets to explore
- Node.js >= 18
- Python 3.11+
- Backend repo: GumusLab/HeatmapBackend
npm install
npm startRuns on http://localhost:3000.
See the HeatmapBackend repo for backend setup instructions.
| Component | Technology |
|---|---|
| Frontend | React, TypeScript, deck.gl, MUI |
| Backend | Django, Django REST Framework |
| Clustering | scipy, scikit-learn, numba |
| AI Chat | OpenAI GPT-4o-mini |
| Docker | nginx + gunicorn + supervisord |
If you use ClusterChirp in your research, please cite:
Rawal, O., et al. "ClusterChirp: A GPU-accelerated Web Server for Natural Language-Guided Interactive Visualization and Analysis of Large Omics Data." arXiv preprint (2026). https://doi.org/10.48550/arXiv.2602.08280
This project is developed by the Gumus Lab at the Icahn School of Medicine at Mount Sinai.
ClusterChirp is freely available for all users. No registration required.