The Intelligent Data Room

A multi-agent web application that allows users to upload data files and "talk" to their data using natural language. The system uses a Planner-Executor agent architecture powered by Google Gemini via LiteLLM proxy and PandasAI.

Live Demo: https://the-intelligent-data-room.streamlit.app/

---

Features

• Data Upload - Support for CSV and XLSX files (Max 10MB)
• Natural Language Queries - Ask questions about your data in plain English
• Multi-Agent Architecture - Planner + Executor agents working together
• Auto Chart Generation - PandasAI automatically generates visualizations
• Context Memory - Remembers last 5 messages for follow-up questions
• Execution Plan Display - See how the AI plans to answer your question

---

Architecture

┌─────────────────────────────────────────────────────────────────────┐
│                         USER INTERFACE                              │
│                      (Streamlit Web App)                            │
└───────────────────────────┬─────────────────────────────────────────┘
                            │
                            ▼
┌─────────────────────────────────────────────────────────────────────┐
│                        ORCHESTRATOR                                 │
│              (Coordinates agents and manages state)                 │
└───────────┬─────────────────────────────────────┬───────────────────┘
            │                                     │
            ▼                                     ▼
┌─────────────────────────┐         ┌─────────────────────────────────┐
│     PLANNER AGENT       │         │        EXECUTOR AGENT           │
│                         │         │                                 │
│ - Analyzes question     │         │ - Receives execution plan       │
│ - Reads data schema     │   ───►  │ - Uses PandasAI + Gemini        │
│ - Creates execution     │         │ - Generates charts (PNG)        │
│   plan with steps       │         │ - Returns answer + visuals      │
│ - Determines chart type │         │                                 │
└─────────────────────────┘         └─────────────────────────────────┘
            │                                     │
            │         ┌───────────────┐           │
            └────────►│ MEMORY MODULE │◄──────────┘
                      │ (Last 5 msgs) │
                      └───────────────┘

Flow

1. User uploads a CSV/XLSX file
2. User asks a question in natural language
3. Planner Agent analyzes the question and creates an execution plan
4. Executor Agent uses PandasAI to execute the plan on the data
5. Charts are auto-generated and displayed alongside the answer

---

Tech Stack

Component	Technology
Frontend/UI	Streamlit
AI/LLM	Google Gemini (via LiteLLM proxy)
Data Analysis	PandasAI, Pandas

---

Quick Start

1. Clone and Setup

git clone https://github.com/yourusername/The-Intelligent-Data-Room.git
cd The-Intelligent-Data-Room
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Configure Environment

Create a .env file:

LITELLM_API_KEY=your_api_key_here
LITELLM_API_BASE=https://your-litellm-proxy.com/v1

3. Run

streamlit run src/app.py

---

Project Structure

The-Intelligent-Data-Room/
├── src/
│   ├── app.py                 # Main Streamlit application
│   ├── agents/
│   │   ├── planner.py         # Planner Agent - creates execution plans
│   │   ├── executor.py        # Executor Agent - runs plans with PandasAI
│   │   └── orchestrator.py    # Coordinates agents and manages state
│   └── utils/
│       ├── memory.py          # Conversation memory
│       └── data_loader.py     # CSV/XLSX file handling
├── exports/
│   └── charts/                # Auto-generated chart images
├── requirements.txt
└── README.md

---

Sample Prompts

• "Create a bar chart showing total Sales by Category"
• "What are the top 5 products by profit?"
• "Show the sales trend over time"
• "Which region has the highest revenue?"
• "Create a scatter plot of Discount vs Profit"

---

License

MIT License