Fama-French models, idiosyncratic volatility, event study

Codes to clean data and construct variables for empirical finance.

Replication package for 'The Anatomy of a Decentralized Prediction Market: Microstructure Evidence from the Polymarket Order Book.' Eight stylized facts on a pre-registered 600-market panel plus a methodological result: feed-inferred trade direction agrees with on-chain ground truth on ~59% of buckets vs ~80% Lee-Ready on equities.

An introduction to database and data management in empirical finance

An introduction to popular databases in empirical finance research.

Cross-sectional Transformer and FFN for stock return prediction and alpha generation. Implements GKX (2020) NN5 replication and MSRR loss (Kelly et al. 2025) for direct portfolio Sharpe optimization. Avg SDF Sharpe 2.05, significant alpha (t=5.34) unexplained by FF5+Momentum.

A toolkit for asset pricing research

A Python tool for extracting stock repurchase program data from SEC 10-Q and 10-K filings

An end-to-end Automated ML pipeline for empirical asset pricing & DJI forecasting. Bridges econometric rigor with modern AI using H2O AutoML. Features include advanced preprocessing (Winsorization, ADF), statistical validation via the Diebold-Mariano test, and model explainability using SHAP values. Optimized for reproducible quantitative research.

From-scratch replication of Loughran & McDonald (2011) — SEC 10-K sentiment analysis with the LM Master Dictionary, Fama-MacBeth regressions on filing-period excess returns.

Research pipeline for building a daily Market Criticism Index and studying its relationship with US market outcomes.

Public release materials for CVE Case Study #3: Multi-Regime Observations Across Fifteen Digital Asset Windows.

End-to-End Python implementation of Mo et al.'s (2025) ACT-Tensor methodology; a tensor completion framework for financial dataset imputation. Implements cluster-based CP decomposition, HOSVD factor extraction, temporal smoothing (CMA/EMA/Kalman), and downstream asset pricing evaluation. Transforms sparse data into dense machine readable data.

Replication of patent-based return predictability using rolling 5-year technology vectors, CRSP returns, Compustat controls, and Fama-MacBeth tests.

This project replicates the connected-firm momentum factor in Ali & Hirshleifer (2020), extends the framework to a quarterly frequency, and applies the same signal construction, portfolio sorts, and factor regressions to an updated 2016–2026 sample.

Project Code of my Master Thesis: The Equity Duration Channel of ECB Monetary Policy Transmission

Quantile Local Projections linking DeFi liquidation shocks to ETH tail risk. Empirical evidence for endogenous market fragility (2021-2025)

Code and data for evaluating informed trading by U.S. Congress members. Employs Fama-French factor models and committee-level jurisdictional mapping to analyze abnormal returns.

Code and reproducibility package for “What Does Deep Hedging Actually Learn? Delta Corrections, Regime Fragility, and Symbolic Distillation.”

CRSP-based analysis of long-run US stock returns, measuring 10-year, 30-year, and full-life holding-period outcomes while explicitly accounting for delistings and survivorship bias.