This project aims to predict customer churn for a major French luxury retail brand. Customer churn is when customers stop doing business with a company. By identifying customers likely to churn, the company can take proactive measures to retain them, thereby increasing customer retention and profitability.
The project is divided into the following main steps:
- Data Engineering: Processing the raw data for exploratory data analysis (EDA) and preparing it for machine learning (ML) model training.
- Visualization: Using Power BI to create visualizations to understand better and interpret the data.
- ML Model: Building an ML model using XGBoost to predict customer churn with an accuracy of 80%.
- Data preprocessing and cleaning.
- Feature engineering to extract relevant features from the raw data.
- Handling missing values and categorical variables.
- Splitting the data into training and testing sets.
- Creating visualizations using Power BI to explore patterns and trends in the data.
- Visualizing critical metrics related to customer churns, such as customer behaviour, purchase patterns, and demographics.
- Building an ML model using XGBoost, a robust gradient-boosting algorithm known for its efficiency and accuracy.
- Feature selection and hyperparameter tuning to improve model performance.
- Evaluating the model using relevant metrics such as accuracy, precision, recall, and F1-score.
- The ML model achieved an accuracy of 80% in predicting customer churn.
- Insights from the visualization helped better understand customer behaviour and identify factors contributing to churn.
- Recommendations and actionable insights to reduce churn and improve customer retention.
Customer churn prediction is crucial for businesses, especially in the highly competitive retail industry. By leveraging data engineering, visualization, and ML techniques, companies can gain valuable insights into customer behaviour and take proactive measures to retain customers and increase profitability.