A modular quantitative finance framework for data collection, analysis, strategy backtesting, and machine learning across multiple markets.
ZVT is a modular quantitative finance framework that provides tools for data collection, persistence, analysis, and strategy backtesting across multiple financial markets. It solves the problem of fragmented financial data and analysis by offering a unified system to handle tradable entities, compute factors, and test trading ideas.
Quantitative developers, algorithmic traders, and financial data scientists who need a structured, extensible framework for market analysis and strategy development.
Developers choose ZVT for its modular design, multi-market data support, and integrated machine learning capabilities, allowing rapid prototyping and testing of trading strategies without managing disparate data sources.
modular quant framework.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Provides a consistent schema for recording and querying stocks, indices, and ETFs across China, US, and Hong Kong markets, as demonstrated with Stock.query_data() examples for different regions.
Includes built-in tools like MaStockMLMachine for training and predicting on financial time-series data, enabling rapid prototyping of ML models with minimal code, as shown in the prediction example.
Emphasizes a modular design with extensible data providers (e.g., Eastmoney, JoinQuant) and schemas, allowing easy integration of new data sources and factors, supported by provider maps in the code.
Offers Dash/Plotly-based UIs for backtesting research and a separate REST API/UI for real-time interactions, facilitating both analysis and live trading workflows, as highlighted in the UI sections.
The project explicitly states it does not guarantee backward compatibility, advising users to upgrade with caution, which can lead to breaking changes and maintenance headaches for dependent systems.
Setting up the full system requires running multiple data runner scripts, deploying a separate frontend UI, and configuring data providers, making it cumbersome for quick starts or small teams.
Real-time market quotes depend on the QMT data source, which requires contacting the author for access, introducing potential lock-in and flexibility issues for users needing alternative feeds.