Alex Trommer

Data Science · University of Michigan · alextrommer@gmail.com · GitHub

About

I'm a data scientist and recent University of Michigan graduate (B.S. Data Science, May 2026) with a focus on machine learning, statistics, and applied analytics. My coursework spans ML systems, data analysis, and statistical inference.

I build sports analytics tools, so far primarily focused on soccer. I'm interested in how statistical methods can surface structure in data that's otherwise hidden in the noise of a match: where to aim a set piece, how to model shot quality, which signals actually predict outcomes.

This site collects the projects I've built and analyzed along the way.

Projects

Ideal Corner Kick Delivery Zones Statistical analysis of 34,000+ corner kicks from StatsBomb open data, identifying which delivery zones produce a meaningful improvement over the 2.9% baseline success rate.
Python StatsBomb Logistic Regression z-test
League Winner Predictor KNN classifier trained on 88,310 player-season records across seven European leagues, predicting whether a forward's team won the league using only goals and shots on target. 95.45% accuracy.
Python KNN GridSearchCV sklearn
xG Dashboard XGBoost expected goals model and interactive dashboard covering the top 5 European leagues, built on ~257,000 shots. Three situation-specific models with isotonic calibration. Refreshes daily via GitHub Actions.
XGBoost Python Railway GitHub Actions