Fractals in Finance

Using fractal analysis for stock price prediction and market analysis.

Overview

This project demonstrates how to apply fractal mathematics to financial data analysis. It calculates the Hurst exponent to determine if a stock price series exhibits persistent trends, mean reversion, or random walk behavior, which can inform trading strategies.

Features

• Hurst Exponent Calculation: Measures fractal dimension of price series
• Trend Prediction: Classifies market behavior based on Hurst value
• Data Visualization: Plots price series with Hurst analysis
• API Integration: Fetches real-time stock data

Data Sources

Alpha Vantage API

• Type: Free financial data API
• URL: https://www.alphavantage.co/
• Features: Daily stock prices, forex, crypto
• Limits: 5 calls/minute, 500 calls/day (free tier)
• How to Use:
  1. Sign up at https://www.alphavantage.co/support/#api-key
  2. Get your free API key
  3. Use in code: api_key = "YOUR_KEY"

Alternative Open-Source APIs

Yahoo Finance (yfinance Python library alternative)

• Use HTTP.jl to fetch from Yahoo Finance endpoints
• Example: https://query1.finance.yahoo.com/v7/finance/download/AAPL?period1=0&period2=9999999999&interval=1d&events=history

Polygon.io

• URL: https://polygon.io/
• Features: Stocks, options, forex, crypto
• Free tier: 5 calls/minute

Financial Modeling Prep

• URL: https://financialmodelingprep.com/
• Free tier available

IEX Cloud

• URL: https://iexcloud.io/
• Free tier for limited data

Installation

1. Clone repo: git clone https://github.com/blakkhatt/FractalFinance.git
2. Activate Julia environment: cd FractalFinance && julia --project

Usage

1. Get API key from Alpha Vantage
2. Edit main.jl and set your api_key
3. Run analysis:

   julia --project
   include("main.jl")
   hurst, pred = analyze_stock("AAPL", api_key)

Hurst Exponent Interpretation

• H > 0.6: Persistent (trending) - likely to continue in same direction
• 0.5 < H < 0.6: Weak persistence
• H ≈ 0.5: Random walk
• 0.4 < H < 0.5: Weak mean reversion
• H < 0.4: Anti-persistent (mean-reverting) - likely to reverse

Code Explanation

fetch_stock_data()

• Uses HTTP.jl to query Alpha Vantage API
• Parses JSON response into DataFrame
• Returns historical daily close prices

hurst_exponent()

• Implements R/S analysis for Hurst calculation
• Fits log-log regression to estimate fractal dimension
• Returns Hurst value between 0 and 1

predict_trend()

• Classifies market behavior based on Hurst value
• Provides trading insights

analyze_stock()

• Orchestrates data fetching, analysis, and visualization
• Saves plot as PNG file

Applications

• Trend Following: Use high Hurst for momentum strategies
• Mean Reversion: Use low Hurst for contrarian strategies
• Risk Assessment: Hurst indicates market efficiency
• Portfolio Optimization: Allocate based on fractal properties

Contributing

Add more fractal indicators, prediction models, or data sources!

Disclaimer

This is educational only. Not financial advice. Past performance ≠ future results.