LLM-Powered Data Analyst

Upload a CSV, describe the analysis you want, and let the LLM generate and execute Python code (pandas + matplotlib) to answer your question.

Features

• Upload any CSV at runtime (no hardcoded data.csv).
• LLM-generated analysis code with safety checks before execution.
• Preview data, see generated code, view resulting table and chart.

Prerequisites

• Python 3.10+
• A Groq API key stored in .env as GROQ_API_KEY=<your_key>.

Setup

1. Create and activate a virtual environment:

   python -m venv venv
   . venv/Scripts/activate    # Windows
   source venv/bin/activate   # macOS/Linux

2. Install dependencies:

   pip install -r requirements.txt

3. Create .env with your Groq key:

   echo GROQ_API_KEY=your_key_here > .env

Run

streamlit run app.py

Then upload a CSV and enter your analysis prompt (e.g., "Show total sales per region as a bar chart").

Live Demo

• Try it in the browser: https://llm-data-analysis-executor.streamlit.app/
• The repository includes a sample sales file data.csv you can upload for quick trials.

How it Works

• app.py: Streamlit UI for upload, preview, LLM invocation, and result display.
• prompt.py: Builds the LLM prompt using the uploaded file path and schema.
• llm.py: Calls Groq to generate analysis code.
• executor.py: Validates and executes generated code with restricted imports.
• schema.py: Reads column names from the uploaded CSV to inform the prompt.

Notes

• Only pandas and matplotlib are allowed in generated code; writing files is limited to output.png.
• Uploaded CSV is saved to a temporary path for the duration of the Streamlit session.