← Back homeSelected work
Projects
Software for exploring sports and game data, with attention to how it is collected, tested, and presented.
Full-stack development · data visualization · testing
Reviewing how a League of Legends match developed. Compare two champions’ gold, creep score, and experience over time, then inspect recorded events within a selected interval.
JavaSpring BootTypeScriptNext.js / ReactTailwind CSSRechartsPostgreSQLFlywayDockerJUnitVitestPlaywrightTestcontainersGitHub ActionsVercel / Railway
Explore live match data or try the built-in synthetic sample. Missing source data stays visibly unavailable.
Machine learning · data engineering · testing
Predicting NBA points, assists, and rebounds from information available before tipoff. A Python pipeline with chronological evaluation, leakage-safe features, and a simple baseline to measure against.
PythonpandasNumPyscikit-learnSQLAlchemySQLiteMatplotlibpytestGitHub Actions
The points model beat the rolling-average baseline in 23 of 28 trailing seven-day windows, with a small overall improvement.