Spaces:
Runtime error
Runtime error
A newer version of the Gradio SDK is available:
6.6.0
metadata
title: Bertopic Gradio
emoji: ๐
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 4.16.0
app_file: app.py
pinned: false
BERTopic Topic Modeling Gradio App
A user-friendly web application for topic modeling using BERTopic with Hugging Face embeddings. Upload a text file and visualize discovered topics with an interactive intertopic distance map.
Features
- File Upload: Upload any text file (.txt) for analysis
- Automatic Document Detection: Intelligently splits text by paragraphs or lines
- Hugging Face Embeddings: Uses sentence-transformers for high-quality embeddings
- Interactive Visualization: Explore topics with a Plotly-based intertopic distance map
- Topic Explorer: Get detailed information about specific topics
- Customizable Parameters: Fine-tune UMAP and HDBSCAN settings
Usage
1. Prepare Your Data
Create a text file where:
- Each line or paragraph is treated as a separate document
- Documents should have at least 3 words each
- For best results, provide 20-50+ documents with varied content
2. Upload and Process
- Click "Upload Text File" and select your .txt file
- Adjust advanced parameters if needed (or use defaults)
- Click "๐ Run Topic Modeling"
3. Explore Results
- Intertopic Distance Map: Interactive visualization showing topic clusters
- Topic Table: Shows topic IDs, document counts, and top keywords
- Topic Explorer: Enter a topic ID to see detailed keyword weights and representative documents