An Anti-churn Model for Real-Time Business

It is a principal factor to understand the business churn. Customer churn, also known as customer attrition, is the measure that tells the number of customers currently associated with the business but will not continue with the business in the future. It is crucial to understand why customers are turning over and by doing so the business strategy can be improved to retain the customer which will result in reducing churn. To develop a realistic solution, we have worked on real-time business data for an Italian company. Initially, we started with different raw datasets. We have performed extensive preprocessing which includes data cleaning, managing missing values, merging different datasets into one, feature selection as well as developing the target variables. Exploratory data analysis (EDA) has been an important process of exploring and understanding raw datasets. Using EDA, we have understood the pattern of different variables and the relationship between these variables and gathered different insights for hypothesis testing and modelling. Using feature selection techniques, we have identified significant features having a potential impact on customer churn as well as developed several classification models such as Random Forest, Classification Trees, Support Vector Machine (SVM), Linear discriminant analysis (LDA) as well as Generalized Linear Model (GLM) to perform prediction about a potential customer will be a churn or not and concluded the discussion by comparing the performance of these models. The goal is to improve the business by identifying patterns and trends in customer churn and providing insights on how to retain customers and improve market performance.