FinHackCN/finhack

What it solves

FinHack is a quantitative finance framework designed to streamline the entire investment research workflow. It provides a unified system for data collection, factor computation, and strategy backtesting, reducing the complexity of building custom quantitative trading systems from scratch.

How it works

The framework integrates several core modules for the quantitative pipeline:

• Data Collection: Includes built-in collectors for sources like Tushare.
• Factor Engine: A computation engine that supports formula-based calculations for factor sets like Alpha101 and Alpha191.
• Machine Learning: Supports multi-process training to maximize server compute power.
• Backtesting: A system that handles A-share market rules (such as T+1 constraints and price limits) and utilizes a dynamic adjustment mechanism to maintain price accuracy and speed.
• Execution: Provides hooks for live trading integration.

Who it’s for

Quantitative researchers (Quants) and financial researchers who need an extensible framework to develop, test, and mine for new alpha factors in various markets.

Highlights

• Extensible Architecture: Users can easily add their own data sources, AI models, and trading strategies.
• Market-Specific Rules: Built-in support for A-share trading constraints.
• High Performance: Uses multi-processing for both backtesting and machine learning training.
• Formula-Based Factors: Native support for calculating complex alpha factors via formulas.