Fantasy Football offers an excellent opportunity to leverage data to predict future performance of players, utilizing machine learning modeling. In this talk, a deep-dive discussion into the code to generate an excellent model, following the Cross Industry Process for Data Mining (CRISP-DM) process will be discussed. An in-depth review of feature engineering, feature selection, hyperparameter optimization techniques and A/B testing of experiments will be discussed. The talk will conclude with a demonstration of MLflow, an open source platform empowering significant velocity enhancements for the machine learning lifecycle. This demonstration will detail how a Data Scientist can generate over a thousand model runs and can efficiently review these model results to identify the optimal data and model to solve the Fantasy Football problem of interest, "Which players should I draft?"
Ben Fowler is a Data Scientist in the Automation & Innovation group of JM Family Enterprises. JM Family Enterprises has consistently been ranked in the Fortune Top 100 Companies to Work For, coming in at #19 in the 2019 list. They have open data science positions, so if you are interested, come talk with Ben.
Doors at 6:30 for pizza, drinks, and networking. Talk will begin around 7:15. I would recommend parking in the Evernia Street Public Parking Garage.