World Embedding: a daily 64-dim representation of the aggregate economic state (1985-2021). Drop-in state vector for asset pricing, macro forecasting, regime detection, and event studies.

Forecasting GDP using MIDAS regressions with mixed-frequency macroeconomic indicators; includes data preparation, model estimation, and evaluation.

Forecasting UK GDP growth using ARIMA on quarterly data, benchmarked against naive forecasts

Interpretable ML framework for salary benchmarking and wage anomaly detection across countries, using macroeconomic parity models (GSB/CDPB), XGBoost, SHAP, and counterfactual analysis.

End-to-end Python implementation of Ma et al.'s (2025) matrix-variate diffusion index models for macroeconomic forecasting. Features α-PCA factor extraction, supervised screening, and ILS estimation for high-dimensional forecasting with preserved structural information.

End-to-End Python replication of Iadisernia & Camassa’s LLM macroeconomic forecasting methodology (ICAIF 2025). Implements: 2,368 synthetic economist profiles, 120,000+ GPT-4o forecasts across 50 European Central Bank (ECB) SPF rounds, a rigorous ablation study with Monte Carlo & binomial hypothesis testing.