Assume that the user is in churn if his/her last event has occurred at least a week ago. Using the event data(export_api) for users with version 1.0.8 onwards and user data(users), segment each user who has not churned yet, to one of the 3 clusters:
- Users who are least likely to churn 0-25%
- Users who are most likely to churn 75-100%
- Other users - 25-75%
The data is stored on MongoDB NoSQL database called ‘data’ and can be accessed at -------. There are three collections in total:
- “export_api” is a collection of documents that store events data
- “users” is a collection of documents that store up-to-date user data
- “script_data” is a collection that is not required in the task below Please see the data definitions and events list here.
The solution should consist of:
- One file with the results of the task that contains two fields:
- amplitude_id - unique id of a user
- The segment the user belongs to
- The code that was used for solution - link to GitHub or archived file
- A pdf file that describes the following points:
- Why the proposed model is a good solution?
- Your steps were taken to achieve the proposed solution