Churn Analysis of the customers of Telecom Industry

日期 : Nov 2016 - Aug 2017

Telecommunication sector in Pakistan is progressing by leaps and bounds. With the rapid growth and increasing competition in telecommunication companies, there are numerous options for the customers to choose from, giving rise to customer churn (Rate of attrition or percentage of subscribers to a service who discontinue their subscriptions to that service within a given time period). This project is aimed at determining the factors for customer churn and designing a model that predicts churn. It required collection of telecom users’ data that underwent data mining techniques to produce meaningful results determining the reasons for customer churn. We took nineteen different parameters into account which affect customer churn, such as: monthly charges, payment methods, contract type, tenure, gender, age group etc. Then different modelling techniques are used on the data to find the best model that gives more accuracy in prediction of churn. The algorithm has been modeled using R as the statistical tool. The project determines the factors of customer loss, making it easier to reduce and eliminate those factors. The predictive model shows whether the customer will churn or not, giving the company ample opportunity to save its customer through proper customer support strategies.