Personalized Recommendation Engine using Data Lake for a Leading American Telecom Operator

About the Client

The client is a major mobile network operator in the US, serving more than 5 million customers across states.

Business Requirement

  • To create a single source of truth & feed 360-degree customer data to a Personalized Recommendation Engine which helps identify the right products for customers based on usage patterns & preferences

Our Solution

  • Designed & created a Hadoop based Data Lake as a common source for all customer data
  • Created Talend based pipelines to source data from EDW, local data marts, etc. at predefined frequencies
  • Created a configurable batch engine to parameterize the running of the pipelines for processing of data

Process Flow

Data Lake Process Flow

Business Outcomes

  • Rapid turnaround time of less than 4 months
  • Single platform for all customer data with up-to-date information and provisions to include additional structured & unstructured data
  • Inherent scalability of the platform to handle data from few hundred Gigabytes to several Petabytes