Jomaric's picture
feat: initial release of machine learning space
b691df4
|
Raw
History Blame Contribute Delete
891 Bytes
---
title: Temporal Topic Modeler
emoji:
colorFrom: yellow
colorTo: green
sdk: gradio
app_file: app.py
pinned: false
---
# Temporal Topic Modeler
An interactive educational computational humanities application designed to map, track, and visualize how conceptual topics and thematic areas evolve and shift over time inside historical document corpora.
### Features
1. **Interactive Timeline Trends**: Visualize topic popularity changes across chronological years or intervals using Plotly stacked area graphs.
2. **NMF Topic Modeling**: Fit a Non-Negative Matrix Factorization (NMF) decomposition model to extract highly coherent latent topics from documents.
3. **Automatic Keyword Definition**: Inspect the top 8 terms defining each topic cluster automatically.
4. **Data Exports**: Download a comprehensive CSV mapping all documents with their detailed topic distribution weights.