Capstone Projects

Predicting Churn of Wireless Customers

Program: Data Science Master's
Host Company: Case Study Project: Data from Kaggle.com
Location: Overland Park, Kansas (onsite)
Student: Ravi Kallepalli

In telecom industry churn represents percentage of subscribers leaving the service provider in a given period of time. In telecom world costs for acquiring a new customer is way more expensive than retaining the existing customers. So all service providers try to reduce churn and its one of the most important KPI’s (key performance indicator) companies emphasize on. This is also one of the key metrics released every quarter during financial reporting to investors. By predicting which customer might leave beforehand, company can take appropriate actions to avoid customer leaving the company. This can save companies millions of dollars from new customer acquisition costs and increased revenue and profits by keeping their existing customers.

“Capstone experience was great. Got to learn lot new things. During the entire course all the data which I worked on was pretty clean and so didn’t have to spend time cleaning the data. But during capstone I got to learn of new data preparation concepts which were not covered during the class. Overall I am very impressed with the program. It had everything I was looking for. Professors were great in their responses, never felt I was learning online. Great experience.”