HR Analytics in Retail: Predicting Employee Churn with Machine Learning

The retail sector is facing operational inefficiencies and high recruitment costs due to the increasing turnover rate. This project focuses on identifying and analyzing the key factors leading to the departure of qualified personnel. The data discussed is a real-life HR analytical case of a company working in the retail industry. Using advanced machine learning algorithms such as LGBM, XGBoost, AdaBoost, and CatBoost, the study seeks to reveal the relationship between variables like education level, city, and age. The project provides actionable insights that can inform strategic decisions by using both qualitative and quantitative data sources. According to the results, the most successful model is discovered as CatBoost. The findings indicate that the employee’s average sales and its coefficient of variation, trends of sales, and age of employee play crucial roles in employee churn. To interpret these, an increase in an employee's sales rates correlates with a higher likelihood of retaining their position. Actions taken in light of the project's findings can contribute to companies predicting employee churn in advance, thereby reducing turnover rates and improving operational costs.