Data Sources

Overview

This section documents the data sources and collection methodologies used across all quantitative research projects.

Cryptocurrency Data

Binance API

• Source: Binance Public API
• Asset: BTC/USDT
• Frequency: 1-minute bars
• Features: OHLCV + microstructure data
• Open, High, Low, Close prices
• Volume and quote volume
• Trade count per minute
• Buy/sell ratio (taker buy vs sell volume)
• Average trade size

Data Quality

• Coverage: 24/7 continuous trading
• Completeness: >99.5% data availability
• Latency: Real-time with <1 second delay
• Validation: Cross-checked with multiple exchanges

Data Processing Pipeline

1. Collection: Automated API calls with rate limiting
2. Validation: Outlier detection and missing data handling
3. Feature Engineering: Technical indicators and microstructure metrics
4. Storage: Efficient CSV format with timestamp indexing

Simulated Data

For research purposes, realistic market data is generated using: - Price Dynamics: Geometric Brownian Motion with jumps - Volume Patterns: Realistic intraday seasonality - Microstructure: Correlated order flow and spread dynamics - Regime Switching: Multiple volatility and liquidity states

All simulated data maintains statistical properties consistent with real market behavior.