Clinical Text Extractor: GPT‑4 Powered Streamlit Dashboard
This app demonstrates how to extract structured information from unstructured clinical notes using OpenAI's GPT‑4 model. It enables both CSV-based batch extraction and manual input, while storing results in a SQLite database and providing rich visualizations.
📊 Project Summary
- Uploads CSVs of clinical notes and extracts diagnoses, medications, and follow-up plans using GPT‑4.
- Includes manual input for one-off extraction scenarios.
- Logs extractions in a SQLite database for traceability and historical review.
- Visualizes key extraction metrics and audit summaries in real-time.
🧰 Tools Used
- Streamlit for UI and dashboard interactivity
- OpenAI API (GPT‑4) for clinical note parsing
- SQLite for structured logging and persistence
- Plotly Express for dynamic visualizations
- Pandas for data preprocessing
📷 Dashboard Screenshots
