中文文档 | English
An intelligent data analysis tool powered by LangChain and LLMs that converts natural language questions into Python code, executes analysis, and provides explanations.
- Multi-LLM Support: Gemini, GPT-4, Claude, DeepSeek, Qwen3
- Smart Code Generation: Natural language → Python analysis code
- Auto Error Correction: Retry up to 3 times with error feedback
- Conversation History: Context-aware multi-turn dialogue
- Dual Interface: Web UI (Streamlit) + CLI
- Chinese Chart Support: Matplotlib with Chinese fonts
- Auto Fallback: Switches models on API quota errors
| Home | Analysis |
|---|---|
 |
 |
- Python 3.8+
- At least one LLM API key
# Clone repository
git clone https://github.com/yourusername/data_analyzer_app_with_llm_agents.git
cd excel_agent
# Install dependencies
pip install -r requirements.txt
# Configure API keys
cp .env.example .env
# Edit .env and add your API keysstreamlit run app.pyVisit http://localhost:8501
python cli_analyzer.py data.csv --llm qwen3Data Analysis:
Analyze sales trends of Clothing over time
Which years did Components outsell Accessories?
Visualization:
Create a pie chart of sales by Category
Plot Sales vs Rating scatter chart
Statistics:
Calculate average Rating by Category
Find sales outliers
User Input → LangChain (Unified LLM Interface)
↓
System Prompt + History
↓
LLM Generates Python Code
↓
exec() in Isolated Environment
↓
Success → LLM Explains Result
Failure → Retry (max 3 times)
├── data_analyzer.py # Core engine
├── app.py # Streamlit web UI
├── cli_analyzer.py # Command-line interface
├── test_analyzer.py # Test suite
├── requirements.txt # Dependencies
└── .env.example # API key template
Edit .env file (at least one required):
GOOGLE_API_KEY=your_gemini_key
OPENAI_API_KEY=your_openai_key
ANTHROPIC_API_KEY=your_claude_key
DEEPSEEK_API_KEY=your_deepseek_key
QWEN_API_KEY=your_qwen_key- LangChain: Unified LLM interface
- Streamlit: Web UI framework
- Pandas: Data manipulation
- Matplotlib: Visualization with Chinese support
Contributions welcome! Please feel free to submit a Pull Request.
MIT License - see LICENSE file
⭐ Star this repo if you find it helpful!