Stock Analysis Project with AutoGen

Varshith Gaddam

Data Analyst

pandas

Python

Investment & Asset Management

Stock Analysis Project

A multi-agent system built with AutoGen to fetch, analyze, and visualize stock data using the yfinance API and Google Gemini API for AI-powered insights.

Overview

This project demonstrates a basic agentic framework that:

Fetches real-time stock data.

Calculates financial metrics (e.g., returns, volatility, moving averages).

Generates enhanced visualizations (price trends and volume).

Provides professional analysis and investment recommendations.

Built as part of a task to explore agentic frameworks, this implementation uses AutoGen with a custom yfinance-based tool, enhanced for practical use.

Features

Data Fetching: Retrieves stock price history, volume, and metadata via yfinance.

Performance Metrics: Calculates total return, volatility, and trend direction.

Visualization: Plots price with 20/50-day moving averages and trading volume in a single figure.

AI Analysis: Uses Google Gemini API (gemini-1.5-flash) for financial analysis and reporting.

Output: Generates a markdown-formatted report with key metrics and recommendations.

Prerequisites

Python 3.10 or higher

Git installed

A Google Gemini API key (free tier available at ai.google.dev)

Setup

Clone the Repository:

Installation

Install Dependencies:

Run the following command to install the required dependencies:

pip install -r requirements.txt

Requirements include:

autogen>=0.2.0

yfinance>=0.2.40

matplotlib>=3.8.0

openai>=1.66.3

numpy==1.26.4

pandas==2.2.2

Configuration

Configure API Key:

Open config/config.py and ensure the Gemini API key is set:

config_list = [
    {
        "model": "gemini-1.5-flash",
        "api_key": "your-gemini-api-key-here",
        "base_url": "https://generativelanguage.googleapis.com/v1beta/"
    }
]

Replace "your-gemini-api-key-here" with your key from Google AI Studio.

Usage

Run the Script:

Execute the script using:

python main.py

Input a Ticker:

Enter a stock ticker symbol (e.g., AAPL for Apple) when prompted.

Press Enter for the default (AAPL).

Type q to quit.

Output:

The script generates a detailed report with:

A table of key metrics.

A stock analysis summary.

A base64-encoded chart for visualization.

Example Output:

AAPL Stock Report

Metric Value Current Price $225.45 Total Return 25.8% Volatility 28.4%

Recommendation: Buy

Price and Volume Charts

[Base64 image data]

Project Structure

stock_analysis_project/
├── agents/
│   ├── financial_analyzer.py  # Data fetching and metrics
│   └── stock_agents.py       # Multi-agent system
├── config/
│   └── config.py            # API configuration
├── utils/
│   └── visualization.py     # Enhanced plotting
├── main.py                  # Entry point
└── requirements.txt         # Dependencies

Enhancements

Multi-subplot charts for price and volume.

Volatility and trend analysis.

Optimized data sampling for efficient visualization.

Markdown-formatted reports for readability.

Limitations

Requires an internet connection for API calls.

Gemini API free tier has rate limits (15 RPM, 1500 RPD).

Base64 charts are text-based; consider saving as files for easier viewing.

Contributing

Feel free to fork this repository, submit pull requests, or open issues for suggestions!

License

This project is unlicensed—free to use and modify as you see fit.

Like this project

Posted Apr 20, 2025

Developed a multi-agent system for stock data analysis using AutoGen, yfinance, and Google Gemini API.

Likes

Views