Private Data Intelligence Engine (PDIE)

A privacy-first, local AI application that validates CSV data quality and explains issues in plain language — without sending data to the cloud.

What it does

• Upload a CSV file via a simple web UI
• Runs deterministic data checks (emails, duplicates, missing fields)
• Uses LLaMA 3 via Ollama (local) to generate human-readable explanations
• No APIs, no external calls, no data leaves the machine

Why it matters

Most data quality tools require cloud uploads or technical expertise.
PDIE is designed for non-technical users who need fast, explainable insights while keeping data private.

Tech Stack

• Python
• Streamlit (UI)
• Ollama (local LLM runtime)
• LLaMA 3
• Rule-based validation + AI explanation layer

Architecture

Streamlit UI ↓ Rule-based Validation Engine ↓ Local LLM (LLaMA 3 via Ollama)

How to run locally

pip install streamlit
python -m streamlit run app.py

## How to run locally
```bash
pip install streamlit
python -m streamlit run app.py
Runs fully offline once dependencies are installed.

Example Use Cases

Marketing lead list validation

CRM data hygiene checks

Sales ops CSV audits

Privacy-sensitive datasets

Privacy

No cloud APIs

No tracking

**## Security & Privacy
- Runs entirely on localhost
- No external API calls
- No credentials required
- No data persistence
- Designed for privacy-sensitive datasets


No data storage**

Runs entirely on localhost

Built With

• python