CSV/Sheet Analyzer with LLM

Overview

CSV/Sheet Analyzer with LLM is a powerful, interactive web application that leverages Large Language Models (LLMs) to analyze, visualize, and extract insights from your CSV or Excel (XLSX) files. With a user-friendly Streamlit interface, you can upload your data, ask natural language questions, and receive instant Python code, visualizations, and AI-generated insights—all without writing a single line of code yourself!

---

Features & Capabilities

• Upload CSV/XLSX Files: Supports both CSV and Excel files for maximum flexibility.
• Natural Language Queries: Ask questions or request visualizations in plain English.
• Automatic Code Generation: The app uses an LLM (Gemini 2.0 Flash) to generate Python code for your query, using pandas, numpy, matplotlib, and seaborn.
• Safe Code Execution: All generated code is executed in a sandboxed environment for security.
• Instant Data Preview: View the first 100 rows of your uploaded data.
• Visualizations & Tables: Get plots, tables, or text results based on your query.
• AI-Powered Insights: For visualizations, the LLM provides concise, context-aware insights about your data and the generated plot.

---

How It Works

1. Upload your data: Choose a CSV or XLSX file to analyze.
2. Preview: Instantly see the first 100 rows of your data.
3. Ask a question: Type a question or request (e.g., "Show the distribution of sales by region" or "What is the correlation between age and income?").
4. LLM generates code: The app sends your query and data context to the LLM, which returns Python code.
5. Sandboxed execution: The code is safely executed. Results (plots, tables, or text) are displayed.
6. Get insights: For plots, the LLM provides additional insights and observations about the visualization.

---

Installation & Running Locally

1. Clone the Repository

git clone https://github.com/harshm2601/SheetMinds.git
cd SheetMinds

2. Install Dependencies

Navigate to the app directory and install the required Python packages:

pip install -r requirements.txt

3. Run the Application

streamlit run main.py

The app will open in your browser at http://localhost:8501.

---

File Structure

app/
  main.py           # Streamlit app entry point
  llm.py            # Handles LLM API calls and code extraction
  sandbox.py        # Securely executes generated code
  utils.py          # Utility functions for file handling and data overview
requirements.txt  # Python dependencies

---

Requirements

• Python 3.8+
• See app/requirements.txt for Python package dependencies

---

Security & Limitations

• Code Execution: All code is executed in a restricted environment, but always review generated code for safety.
• API Key: The LLM API key is hardcoded for demonstration. For production, use environment variables or a secure vault.
• File Size: Large files may impact performance.

---

Credits

• Built with Streamlit, Pandas, Matplotlib, Seaborn, and Gemini LLM.

---

License

This project is for educational and demonstration purposes only.