Hi, I'm Mason,

Business Analyst @ Madison Gas & Electric.
DS and CS @ UW-Madison.

Advanced professional experience in MLOps, automation, and data viz.
Empowering teams to make data-driven decisions.
Writing about sports and models, occasionally.

about

As a Business Analyst at MGE since January 2026, I’ve owned projects from start to finish–enhancing, modernizing, and automating key processes across multiple teams with ML systems, Python, DuckDB, Microsoft Power Automate, and Power BI. I specialize in enabling customer billing and customer interaction through the development of new tools and procedures.

At UW-Madison, I took coursework with an emphasis on machine learning engineering, data systems, and statistical analysis. This includes coursework in neural networks, advanced statistical analysis, probability, and general machine learning applications. In addition, multiple semesters also included intensive project-based work with Java, C, R (tidyverse, ggplot, ranger), and React Native.

You’ll also find me writing and creating models at FiftyFortyNinety–a sports analytics project I started in early 2026. I cover sports and leagues ranging from college football and FCS to soccer and the EPL, MLS, and USL. Some notable projects include my work projecting the final 2026 NCAA MBB bracket based on real-time results and analyzing the 2026 FIFA World Cup group stage, match-by-match.

portfolio

2026 NCAA March Madness Bracket Projections
End-to-end ETL pipeline taking in real-time standings and NCAA MBB results to full 68-team bracket and bubble projections with seeds and matchups for the 2026 NCAA MBB tournament.
2026 FIFA World Cup Group Stage Projections
Projected probabilities for each team's progression into the knockout rounds and for each matchup in the 2026 FIFA World Cup group stage, with team form and ELO.
2025-26 NCAA Men's and Women's Basketball ELO Dashboard
R and Python ETL pipeline and custom-built D3.js visualization for team rankings for all 365 and 363 D1 men's and women's teams and their respective conferences.
Estimated Possession Value and Deriving Expected Threat in Basketball Play-by-Play Data
Multi-model Random Forest and logit pipeline to assign weighted possession value (xT) based on beginning and end state metadata throughout the 2024-25 NBA regular season.
Estimating Completions Prevented Above Replacement (CPAR) with Multi-Level GLM + RF Model
Analysis of frame-by-frame player data as part of 2025 NFL Big Data Bowl to assess defensive coverage of receiver routes via a multi-level GLM and Random Forest model.
USL Championship and League One Team and Player Performance Analysis
Various ad-hoc summaries of team performances and specific player contributions, including analysis of new players, matchup probabilities, and playoff races.
PWHL API
Open-source Python wrapper for retrieving women's hockey player and team data from PWHL.com.
Explore Trempealeau County — Mobile Application and Desktop Content Management System
Full-stack, multi-product solution using React Native, FastAPI, Maplibre, and Amazon S3 for map-based mobile application and content management web app.
Fellow — AI-Powered Feedback and Text Editor
LLM-powered writing platform to connect writers with others, built with Vue.js, Firebase, and Gemini.
Identifying NBA-Caliber G-League Players with Random Forest Similarity, K-Means Clustering
Analytical investigation of G-League player performance extrapolated to NBA potential, using Random Forests and neural networks with players' career numbers and physiological data to project major league success.