Research/Projects

Live, interactive world map tracking personal studies in physical and human world geography. After frequent 7x5 inch notebook (pen, paper) page write-ups, the corresponding countries, territories, and maritime chokepoints are logged; covered countries are shaded green, and chokepoints are pinned. The map automatically updates from a CSV log through a Python and GitHub Actions pipeline.

This map makes no strategic or analytical claims; it is simply a record of disciplined study and an ongoing exercise in data pipeline design.

End-to-end pipeline building the Swing Output Value (SOV) metric, analyzing park-level pitch mix, swing outcomes, and quality over 175,000 swings from 2023–2025 seasons. Includes Statcast pulls, per-season CSVs, downstream SQL analytics, and 7 reproducible notebooks.

An analysis of all 2,586 NCAA Men's Basketball Tournament games from 1985 - 2025, examining which final score digit combinations appear most frequently and why the board is less random than it appears. Structure under randomness.

A supervised regression pipeline trained on just over 6,000 tabular listings to predict used-car sale prices. Workflow performs thorough cleaning and feature engineering on listing attributes, then trains feedforward neural networks using *TensorFlow* with regularization and learning-rate scheduling.

A mini-project focused on capturing and analyzing my daily keystrokes across lecture notes, emails, and everyday use, transforming raw input into structured visualizations through a custom-built data pipeline to better understand how data is stored and used in early machine learning workflows.