Minimizing Attrition Rate Through Analysis of Financial Habits
Subcription products are a main source of revenue for companies across all industries. Since most of these products are in the form of one size fitting all, there are always people who would unsubscribe to the product if they don't like it later on. This project consists of a dataset from a fintect comany which contains 31 different features of 20,000+ customers and the boolean of whether they unsubscibed or not. We clean the dataset, select appropriate features and build a model capable of predicting if a user will unsubscribe or not with an accuracy of 63.2%.
Codebase: Python 3.8.1 (Repo contains two ipynb files).