Churn Analysis in Polythene and Plastic Manufacturer

Abstract

Data mining is the process of searching through large amount of historical data in order to identify patterns and predictive information in data. Data mining is a valuable tool in the current business world which could provide vital information that can help growth of current business. Information is a vital asset to a company and data mining tool could provide a competitive edge over other companies. Polypak Secco (pvt) Ltd. at present is challenged with customer churn management as they have no clear understanding of what causes churning in the company. The company has identified it is far more profitable to investigate the root causes for why customers are leaving. To address the problems of the company and to cater for their requirement this research was developed. This thesis is devoted on the analysis of the company’s historical data as an attempt to predict customer churning. The main aim is on the application of a suitable data mining technique for the identification of churning and based on historical data to discover patterns or the effect of certain attributes on churning. The research was carried out following the CRISP-DM methodology throughout the project cycle. Research resulted in discovering hidden churn factors from the historical data that was thought to be unworthy.