From Dipanjan Biswa on the RudderStack blog:

There are a lot of blogs out there on building churn prediction models using scikit-learn, R, or other advanced ML toolkits. However, most of them require strong engineering as well as data science skills. That said, it is now possible to create robust models using SQL and basic knowledge of data science by leveraging the tools that companies like Google have developed to democratize machine learning.

In this post, we will use Google’s BigQueryML together with the clickstream data collected and delivered using the RudderStack platform. This data represents events, such as slot machine spins, associated with a mobile casino game. The volume of the data is typical for such scenarios, and as we will see, performing churn analysis is both easily accessible and efficient without having to spend a fortune on the infrastructure.

Try Rudderstack Cloud for free