This product focuses Customer Churn Prediction which involves identification of customers who are likely to leave which helps to approach them with marketing offers with a few features. The features of the product includes showing churn probability for existing customers with high/medium/low risk, ranking of customers according to highest churn, historical churn visualisation plots, churn location maps and correlation plots between attributes affecting the churn.
This system is implemented in a way to handle big data, using HDFS to store, 360 degree customer data, Apace Spark for fast data processing, Spark MLib for the prediction, Elastic search and Zepplin/D3 charts for visualisation of the results. The system also allows for easy integration with existing client infrastructure. This product is initially planned to be implemented in the Telecom sector and later could be adapted to other sectors.
The success of this project involved strategising how effectively to put together all of the brand elements to ensure a cohesive and smooth user experience. Research was carried out and brand guidelines were made to ensure consistency throughout the brand’s online presence.