Analysis of Churn Factors of Users of the Online Clothing Store

The online clothing sales market has grown rapidly in recent years. In Russia, there is high competition between existing players, so the problem of user retention is now very relevant. In this paper, the topic of the outflow of users of an online clothing store was discussed in detail. The tasks set at the beginning are fully completed: 1. In the first chapter, an analysis of the relevant literature was carried out and conclusions were drawn about the "nature" of the phenomenon (outflow). 2. In the second chapter, a set of factors was formed that could potentially explain the outflow of the user. 3. Further, the preparation of data, its purification, and exploratory analysis were successfully completed. 4. 4 different machine learning algorithms were trained. Based on the binary classification metrics, the XGBoost classifier was chosen as the base model for further analysis. 5. Using the SHAP method, an analysis was made of the relationship between a set of factors and a target outflow variable. Identified patterns in the data.