Predicting Customer Churn Model + Business Use Cases


Hi all, we're back with another Meetup taking place early September!

Most Marketing and Sales departments understand that advanced analytics can help detect, anticipate, and mitigate customer churn, but the steps to actually accurately predicting churn are often unclear.

End-to-end, from raw data to production, how can a sales/marketing department deploy a churn prediction model?

Along with Keyrus ( we will break down a few use cases of Advanced Analytics for Churn Prediction.

First Talk:

How to Predict Churn with Open Source Technologies
Leo will look at how a UK telecom provider is trying to reduce churn by aggregating customer information using Open Source technologies. In this talk, you'll see how to extract, clean and combine data to make predictions and deploy into production.

Speaker Bio: Leo Yorke
Leo ( is a Senior Analyst at Keyrus UK. He has extensive experience using Python, R and QlikView to develop applications to help clients quantify if and when a customer is likely to leave.

Second Talk: How to Build and Deploy a Churn Management System

Using real-life examples from digital services and energy companies, Vincent ( will show us how to build an deploy a Churn Management System.
From real-time metrics that help customer service agents adapt their message on the fly to batch predictions that drive personalized marketing campaigns, we'll go through two client uses cases to explore how Dataiku ( helps companies address the risk of attrition.

Speaker Bio: Vincent de Stoecklin
Vincent ( is Head of Partnerships at Dataiku (, he has had the opportunity to work with large consulting companies and technical partners on many different big data use cases. His expertise ranges from banking and telco to retail and manufacturing.


This meet-up is for:

- Data geeks of all ages willing to discover how Churn models can be deployed with innovative technologies

- Data professionals from all horizons (any role, any sector)

As anyone who has attended our meet-ups before will testify, you can expect excellent talks, discussion, and complimentary pizza and beer for everyone!


If you haven't completed our 30-second long/4-question survey ( take the time to help us make this Meetup group the best!


Want to collaborate with us to uplift the data community? Get in touch (!