🤖📊 CSV Intelligence — AI Agent Chatbot

Chat with your CSV data in plain English. No SQL. No formulas. Just ask.

---

What Is This?

CSV Intelligence is an AI-powered chatbot that lets you have a full conversation with any CSV file.

Upload your data, ask a question, and the agent:

• Writes and executes Python code against your CSV on the fly
• Generates interactive charts (bar, pie, line) rendered in the browser
• Returns a written summary explaining the insight in plain English

No data science skills needed.

---

Demo

Ask a question	Get a chart + summary
"What are the top 10 products by price?"	Bar chart + ranked list
"What % of items are Large vs Small?"	Pie chart + percentage breakdown
"Show me the distribution of stock values"	Line chart + stats summary

---

Features

• 🧠 AI-powered analysis — DeepSeek V3 reasons step-by-step to answer complex data questions
• 📊 Auto visualisation — generates Chart.js HTML charts tailored to the data
• 🔁 Agent caching — the CSV agent is built once and reused, keeping responses fast
• 🎨 Modern UI — dark gradient theme, live progress steps, sidebar CSV preview & stats
• 💬 Sample questions — clickable chips to get started immediately
• 🛡️ Resilient parsing — json-repair handles imperfect LLM JSON output gracefully
• 🗑️ Auto cleanup — temporary files deleted on exit

---

Tech Stack

Layer	Technology
UI	Streamlit
AI Agent	LangChain — ReAct CSV Agent
LLM	DeepSeek V3 (deepseek-chat)
Data	Pandas
Validation	Pydantic
Package Manager	uv

---

Project Structure

├── ModernChatbot.py   # Modern UI — start here
├── Chatbot.py         # Original UI
├── csv_agent.py       # LangChain ReAct agent (reads & analyses CSV)
├── plotter.py         # DeepSeek formatter (generates charts + summaries)
├── pyproject.toml     # Dependencies
└── .env               # API keys (not committed)

---

Getting Started

1. Clone the repo

git clone https://github.com/kamalviewcode-spec/Project_6_AI_Agent_CSV_Visual.git
cd Project_6_AI_Agent_CSV_Visual

2. Add your DeepSeek API key

Create a .env file in the project root:

DEEPSEEK_API_KEY=your_key_here

Get your key at platform.deepseek.com/api_keys

3. Install dependencies

uv sync

4. Run the app

streamlit run ModernChatbot.py

Your browser will open at http://localhost:8501

---

How It Works

User question
     │
     ▼
LangChain ReAct Agent
  → writes pandas code
  → executes against CSV
  → returns text answer
     │
     ▼
DeepSeek Formatter
  → generates HTML chart
  → writes summary
     │
     ▼
Streamlit UI
  → renders chart in iframe
  → displays summary

---

License

This project is licensed under the MIT License — see the LICENSE file for full details.

Free to use, modify, and distribute.

---

Built with LangChain · DeepSeek · Streamlit · Python