This is a project I did back in Summer 2018. I downloaded datasets from Riot official websites. By using linear regression model on the dataset we try to analyze champion choices and its effect on winning.
This pretty much the first large scope project I did outside course works. It teaches me basic data analysis and data pre-processing, regression and linear programming techniques plus some basic knowledge of machine learning, including test training datasets, overfitting vs underfitting, etc.
linearReg.py is used to build the model. "tester.py" is used for testing prediction accuracy of the model. The model generated over 75% on predicting winning team based on champion choices and other in game statistics using a regression model.