To see my previous podcast on sports analytics, take a look at Triple Play Analytics.
For the past few years, I have been doing extensive research in sports analytics. From analyzing baseball pitchers to creating new metrics for NHL goalies, I have been working on several projects. Here, you’ll find a brief explanation of each project as well as any related links to them.
NHL Goalie Starter Model
In an attempt to determine if a NHL goalie was performing at a starter’s level, I created a machine learning model to determine if a goalie was a starter. In order to keep the definition of starter fair, I determined that a starter is a goalie who starts at least 40 games in a season. The machine learning model is a neural network classifier which takes in various statistics and returns a probability. If the probability is greater than 0.5, then the goalie is deemed a starter. Notable analyses have been done on the Pittsburgh Penguins goalies – Matt Murray and Casey DeSmith.
Links:
- Has Matt Murray Been Performing At Starting Goaltender Caliber?
- GitHub Repo of Main Analysis
- A Revisit: Are Matt Murray And Casey DeSmith At The Starting Goaltender Caliber?
NHL Shots per Goal Statistic
In the NHL, advanced statistics are beginning to be used. I have been researching and analyzing a new statistic that should serve as a “strike-rate” for NHL goalies and skaters.
Last Update: January 8th, 2019
Links:
- GitHub Repo of Main Analysis
- PDF Write-Up of Analysis
- Interactive Website of Analysis
- The Penguins’ Perfect Goaltending Predicament
NHL Goalie and Skater Heat Maps
In this analysis, I am looking at “heat maps” or specific goalies and skaters. These heat maps will show shot volume and scoring likelihood. This is a stepping stone project for a large project: using machine learning to determine the best scoring zones for specific skaters, and using machine learning to determine the weaknesses for goalies.
Last Update: January 31th, 2019
Links:
- GitHub Repo of Main Analysis
- PDF Write-Up of Analysis
- Interactive Website of Analysis
- Jake Guentzel’s Evolution Into An Elite Scoring Winger
NHL Shot Based Goal Scoring Model
This project is aimed to determine if it is viable to use machine learning in determining if a shot will be a goal.
Last Update: February 7th, 2019
Links:
- GitHub Repo of Main Analysis
- Interactive Website of Analysis
- Modeling the Penguins’ Shot Selection
- Triple Play Analysis Podcast: Sports Model Discussions
NHL Shot Type Effectiveness Model
This project is aimed at determining how effective different types of NHL shots are using machine learning.
Last Update: February 24th, 2019
Links:
- GitHub Repo of Main Analysis
- Interactive Website of Analysis
- Triple Play Analysis Podcast: Sports Model Discussions
MLB Pitcher Precision Analysis
This project is aimed to help quantify what precision is to MLB pitchers.
Last Update: May 12th, 2019
Links: