A portfolio prototype demonstrating how a manager or director can query structured business data through a simple conversational interface — no SQL, no dashboards, no complexity.
This is a local demo of an internal data assistant for small and medium businesses. Instead of opening spreadsheets or navigating dashboards, users type plain-English questions and receive instant answers based on the company's structured data.
No LLM. No API calls. No cloud dependency. Just Python, pandas, and Streamlit — clean, fast, and fully offline.
The assistant answers business questions like:
- Which region had the highest revenue?
- What was the best month by profit?
- Compare revenue between March and August
- Show me a summary for the North region
- How did July perform?
- What is the average profit?
data_assistant/
└── demo_03_internal_ai/
├── data/
│ └── business_data.csv ← Fictional sales dataset (48 rows)
├── app/
│ ├── app.py ← Streamlit interface
│ └── logic.py ← Question parsing and data logic
├── screenshots/
├── requirements.txt
└── README.md
cd demo_03_internal_ai
pip install -r requirements.txt
python3 -m streamlit run app/app.pyOpens at http://localhost:8501
| Tool | Purpose |
|---|---|
| Python 3.12 | Core language |
| pandas | Data loading and analysis |
| Streamlit | Web interface |
This prototype is designed as a proof-of-concept for:
- Internal analytics assistants
- Director and manager dashboards
- Business Q&A tools
- Natural-language access to structured data
Portfolio project — built to demonstrate the concept of conversational access to business data.