This project involves using historical data scraped from the public NHL API to make goal predictions for players on the Maple Leafs. Click here to see these models in action, as well as browse through some data visualizations of various statistics. You will be able to make your own predictions of goal success in a simulated game state!
I was introduced to hockey and the NHL about 3 years ago and after some time of just watching games, I began to wonder about how data might be used to make predictions about game plays. Luckily for me, the NHL actually exposes their API to the public and it contains a plethora of data (including data about every single play of every single game!). So I decided to try and use that data and implement two types of machine learning models: binary classification for predicting goal success and regression for predicting the number of goals scored.
Used parameters such as play location and player statistics to predict whether or not a certain shot will reach the inside of the net and gain a point for the Leafs. Random Forest and Support Vector Machine models were implemented.
Used parameters such as player and game statistics to predict how many goals a player would score during a specific game. Elastic-Net and Multi-Layer Perceptron regression models were implemented.