Welcome!
This project was a summer-long endeavour where I learned data analytics and predictive modeling in the context of sports data, ultimately culminating into creating my own XGB model predicting the winner and loser of NHL games based on data from games across all teams with data ranging back to as far as 2008.
In this project, I sourced, cleaned, combined, and processed my data into a format that can be given to an XGB model, and then optimized my model through hyperparameter tuning via a bayesian optimization, which culminated into a model acheiving ~56% accuracy across the '23-'24 NHL season.