What problem does it solve?
End-to-end pain points in quantitative financial research are eliminated by providing a single, repeatable pipeline that handles data ingestion, cleaning, panel modeling, signal generation, and backtesting so analysts avoid manual wrangling across tools and formats.
Core Features & Use Cases
- Reliable data ingestion: Fetch A‑share, fund and macro data with AKShare and fall back to yfinance when needed, with retries and column mapping to a unified schema.
- Standardization & factor engineering: Normalize date/code/price columns, handle missing values, compute rolling factors (momentum, volatility, size) and produce MultiIndex panel datasets for modeling.
- Panel regression & model selection: Run panel OLS with entity effects, cluster-robust errors, model selection guidance and save regression summaries and coefficients for interpretation.
- Signal generation & backtesting: Convert significant factor directions into per-date trading signals and run backtests via backtrader with recommended analyzers and performance reports.
- End-to-end automation: Orchestrated MVP script to execute fetch → prepare panel → run model → generate signals → run backtest and produce persistent CSV/JSON/text outputs for reproducibility.
Quick Start
Run the end-to-end pipeline to fetch A-share data for 000001 and 000002 from 20210101 to 20231231, build panel factors, run a panel regression to extract significant factors, generate trading signals for 000001, and execute a backtest with performance report.
