UCLA

The purpose of this project is to understand the main factors that drive customer tenure within auto insurance industry for six or more years. The analysis is based on three years of the J.D. Power Auto Insurance survey data. For the analysis, multiple binary machine learning algorithms were implemented and measured to classify whether customers would stay with the same insurer for more than six years. Random forest was found to be the most robust model as compared to logistic regression, decision trees, and xgboost.