Description

Used scikit-learn and my library dstricks to analyze, preprocess, and predict whether free customers will convert to paid customer. Also, viewed feature signficance of each variable and determined that time spent on website, first interaction, page views per visist, profile completed at medium, and age were the most significant features of our Decision Tree classification model. I was able to get an F1 Score of 89% on no conversion, and 79% on a conversion to paid since precision was a bit lower, which would be the correctness of our positive predictions (correctly identified convserion / total conversions identified as an actual conversion). Recall would be correctly identified conversion / total actual conversions.