Michael Pilger · Sports Analytics & Data Science
I build quantitative tools for the sports questions I care about.
My work sits at the intersection of sports, statistics, and software—from MLB player valuation and league analytics to athlete-performance data and applied machine learning. I am most interested in models that are testable, interpretable, and useful outside a notebook.
About
Notre Dame data science student focused on predictive analytics, sports modeling, and quantitative tools that connect analysis to decisions.
Biography
Michael Pilger is a Master of Science in Data Science student at the University of Notre Dame. While much of his time is spent studying, working on campus, playing in his Wiffleball league, and chasing Notre Dame's intramural sports appearance record, his primary academic interest is predictive analytics and results-driven data science.
He combines a longstanding interest in sports with statistics, machine learning, simulation, and software development to build models that forecast outcomes and quantify performance, from student-run competitions on campus to major collegiate and professional leagues. Michael hopes to build a career creating quantitative tools that turn complex data into clearer decisions while continuing to pursue independent sports research.
Across his work, he aims to build models that are rigorous enough to trust, transparent enough to understand, and practical enough to use.
- Education
-
University of Notre Dame
M.S., Data Science · 2025-May 2027 University of Notre Dame
B.S., Computational Math & Statistics; Economics · 2021-2025 - Experience
-
Varsity Sports Performance Analyst
University of Notre Dame · Sep 2023-Present AI Training & Evaluation Contractor
Mercor / Handshake · Apr 2025-Present Data Science Assistant & Intern
Danu Analítica · 2026 Sports Management & Data Analytics Assistant
VOLO Sports · 2023-2025 - Research interests
- Sports analytics, player and team valuation, forecasting and simulation, statistical modeling, machine learning, computer vision, and data-product development.
Research and analytical projects
Projects range from original sports research to production data systems and applied machine-learning work.
Window-Adjusted WAR (wWAR)
A prospective MLB player-value framework that holds a player's projected WAR fixed and estimates how efficiently those wins convert into championships for each franchise over a five-year competitive window.
Football Prediction Model
An NFL and NCAAF forecasting system built around leakage-safe pregame features, dynamic team ratings, calibrated win probabilities, projected scores, and transparent out-of-sample performance tracking.
ND Wiffleball Analytics & Operations
A production league platform combining standings, schedules, player statistics, projections, awards, interactive tools, commissioner workflows, captain scheduling, and a persistent historical database.
GeoGuessr Continent Classification
A transfer-learning image-classification project using a pretrained ResNet18 to infer continent from street-level geolocation imagery across five classes.
Applied Mathematical Algorithms
Python implementations spanning sorting and selection, maximum-probability paths in Markov chains, and constrained shortest-path graph search.
Latent-Variable & Nonlinear Modeling
Applied analyses using factor models, structural equation modeling, generalized additive models, diagnostics, and reproducible statistical reporting.