Objective:

Given a list of customers, their transaction history, and a segmentation period, split the customers into 4 segments: Active, Inactive, New and Undefined.

Installation:

This project is built using sbt . It can be imported in an IDE as a sbt project. The language used is scala. Make sure to add scala support in the IDE when importing. You can compile and test with sbt cli using: sbt test
This exercise uses Spark to compute the different segments. All dependencies are handled by sbt and should be resolved without any problem.

Definitions:

• endDate: after this date, transactions are not taken into account.
• segmentation Period (P): period of time ending at endDate, defined by the number of days in the period
• Activity Segments:
  • ACTIVE: the customer have transactions before P and one or more transactions in P
  • INACTIVE: the customer have transactions before P but no transactions in P
  • NEW: the first transaction for the customer is in P
  • UNDEFINED: The customer has no transactions before the end date

Input parameters:

The main method expects 4 parameters:

1. dataPath: this should be projetctDir/data
2. outputDir
3. endDate: yyyy-MM-dd
4. segmentationPeriod

Running the job

• When running the job from an IDE, make sure you include the "provided" dependencies in the classpath.
• To run using command line, start sbt shell then execute run <param> <param> ...

To Do:

You need to implement the methods in src/main/scala/co/ogury/segmentation/SegmentationJob.scala to do the following:

• load the transactions and customers from the data directory using spark.
• compute the corresponding segment for each customer.
• save the customerId,segment couples as a csv file.
• example output:

    customer-id;activity-segment
    42;active
    51;inactive
    64;undefined
    7;new
    747;active

• We have provided a unit tests for 2 segmentation cases. We recommend that you implement the remaining cases in src/test/scala/co/ogury/segmentation/SegmentationJobSpec.scala in order to validate you work.
• run the job with endDate = 2017-12-31 and segmentationPeriod = 365
• commit your changes.
• Add the whole project files and results to an archive and send it by email.