This talk will detail a recent churn prediction project at GSN Games, starting from exploratory data analysis and model development all the way to A/B testing and deploying the churn intervention feature. We will cover some of the successes as well as mis-steps we had along the way.Topics covered include: Defining churn and what this means for your model. Ideas for exploratory data analysis to understand and visualize churn. Developing, diagnosing, and evaluating models for churn prediction. Ideas for churn interventions. A/B testing a pilot churn project. Deploying a churn intervention as an ongoing feature in a game.
Director, Data Science
Dennis O’Brien is Director of Data Science at GSN Games where he is responsible for the full stack from data engineering, data analysis, and predictive analytics. His experience in the game industry extends more than 15 years including Lead Game Programmer at JellyVision Interactive, Software Engineer at LeapFrog Enterprises, and Principal Software Engineer at Electronic Arts. He studied Physics as an undergraduate at the University of Chicago and Computer Science as a graduate student at the University of Illinois at Chicago and worked for many years doing scientific programming, experimental design, and data analysis in Geophysics and Astrophysics.