Systematic Alpha Lab

AI-assisted systematic alpha research platform for factor research, signal generation, alpha evaluation, portfolio construction, and research automation.

Status: active flagship project. The implemented data, factor research, and alpha construction modules have been consolidated into this repository under src/systematic_alpha_lab/. The older step repos remain useful implementation history, but this repo is now the public center of gravity.

Simple Research Workflow

Data Ingestion
    -> Factor Cleaning
    -> Factor Evaluation
    -> Alpha Construction
    -> Risk Management
    -> Portfolio Construction
    -> Backtesting and Attribution
    -> Research Memo Generation

For a credential-free first run, use the synthetic workflow:

python examples/synthetic_equity_alpha_demo.py

That demo runs:

Synthetic prices -> simple factors -> factor analytics -> combined alpha -> IC/IR/turnover

Implemented Platform Layers

Layer	Package	Status	Purpose
Data Pipeline	systematic_alpha_lab.data_pipeline	Implemented	Ingest, transform, quality-check, and serve price/fundamental/macro datasets
Factor Research	systematic_alpha_lab.factor_research	Implemented	Build cross-sectional equity factors, clean signals, run IC/IR and decile diagnostics
Alpha Construction	systematic_alpha_lab.alpha	Implemented	Purify signals, combine thematic composites, evaluate alphas, run walk-forward weighting/ML experiments
Portfolio Construction	systematic_alpha_lab.portfolio	Planned	Turn alphas into constrained weights
Risk Model	systematic_alpha_lab.risk	Planned	Estimate exposures, covariance, limits, and risk budgets
Backtest / Attribution	systematic_alpha_lab.backtest	Planned	Evaluate portfolio returns, costs, turnover, drawdown, and attribution
Agentic Research	systematic_alpha_lab.agents	Planned	Generate hypotheses, run experiments, evaluate outputs, and draft research memos

What This Demonstrates

• Point-in-time data handling for market, fundamental, macro, and benchmark inputs.
• Cross-sectional factor engineering across momentum, value, quality, growth, liquidity, risk, size, forensic, and analyst themes.
• Factor diagnostics: rank IC, IC IR, decile spreads, long-short PnL, Fama-French regression, coverage stats, and rolling analytics.
• Alpha construction with purification against sector, beta, size, and style controls.
• Walk-forward weighting engines: equal weight, IC weighting, MLR, Bayesian shrinkage, GMV/MVO, and regularized ML pods.
• A path toward AI-assisted systematic research: hypothesis agent, experiment runner, evaluation agent, and report generator.

Repository Layout

systematic-alpha-lab/
  src/systematic_alpha_lab/
    core/                # shared dataclasses for research artifacts
    workflows/           # high-level runnable workflows
    data_pipeline/       # data ingestion, transformation, quality checks
    factor_research/     # factor library, cleaning, diagnostics, composites
    alpha/               # alpha purification, weighting, evaluation
    portfolio/           # planned
    risk/                # planned
    backtest/            # planned
    agents/              # planned
    reporting/           # planned
  config/
    datalist.yml
    factors/config.json
    alpha/config.json
  docs/
    architecture.md
    roadmap.md
  examples/
  tests/

Quickstart

Install in editable mode:

python -m pip install -e ".[dev]"

Run the lightweight import check:

python examples/consolidated_import_demo.py
python examples/synthetic_equity_alpha_demo.py
pytest

Credentials are not committed. For live ingestion, create config/credentials.yml or config/credential.yml locally with Alpha Vantage and WRDS credentials.

Configuration

• Data defaults: config/datalist.yml
• Factor cleaning and composite definitions: config/factors/config.json
• Alpha construction and model settings: config/alpha/config.json
• Default data root: ../data, matching the existing local research artifact layout.

Legacy Component Repos

These repos are being consolidated into this flagship codebase:

Legacy repo	Consolidated package
quantlab_step1_data_ingestion	systematic_alpha_lab.data_pipeline
quantlab_step2_factor_research	systematic_alpha_lab.factor_research
quantlab_step3_alpha_library	systematic_alpha_lab.alpha

Documentation

• Architecture
• Roadmap
• Research Workflow