An AI-enabled Market Access analytics prototype that combines LLM-powered natural-language SQL generation with SQL and Pandas-based analytics to analyze the synthetic healthcare claims data.
The application allows users to ask business questions in natural language such as:
"Which payer has the highest rejection rate?"
The system uses a Groq LLM to translate the question into SQL, validates the generated query and executes it against the claims dataset to return analytical insights.
LIVE PROJECT LINK : https://market-access-analytics.streamlit.app/
Market Access analytics teams work with large datasets involving:
- Payers
- Health plans
- Drugs
- Claims
- Formulary tiers
- Prior authorization
- Step therapy
- Quantity limits
- Rejection reasons
- Distribution channels
This project demonstrates how AI + traditional analytics can be combined to make this type of analysis easier and faster.
- Interactive Market Access dashboard
- Claims data exploration
- Natural-language-to-SQL analytics
- Payer rejection analysis
- Rejection reason analysis
- Anomaly detection
- REST API endpoints
- Docker containerization
- GitHub Actions CI
User
|
v
Streamlit UI
|
Natural-Language Question
|
v
Groq LLM
Llama 3.3 70B
|
v
Generated SQL
|
v
SQL Validation
|
v
SQLite / Claims Data
|
+------+------+
| |
v v
SQL Pandas
| |
+------+------+
|
v
Analytics Result
|
v
User
---
# REST API Architecture
Client / Application
|
v
Flask API
|
+----------------+----------------+
| | |
v v v
/health /ask /report/*
| | |
| v |
| Groq LLM |
| | |
+----------------+----------------+
|
v
SQL / Pandas Analytics
|
v
Claims Dataset