Spark - Couchbase Market Basket Analysis

This repository is companion code for a blog post on our site: http://blogs.avalonconsult.com/blog/big-data/combining-operational-and-analytical-big-data-using-couchbase-and-spark-a-market-basket-analysis-example/.

This project is an all in one environment that setups a Vagrant machine with Couchbase and Spark installed. And has a Spark process that will generate some basic Market Basket Analytics in the form of product recommendations.

Sample data taken from:

Brijs T., Swinnen G., Vanhoof K., and Wets G. (1999), The use of association rules for product
assortment decisions: a case study, in: Proceedings of the Fifth International Conference on
Knowledge Discovery and Data Mining, San Diego (USA), August 15-18, pp. 254-260. ISBN:
1-58113-143-7.

Can be found under the 'retail' link here: http://fimi.ua.ac.be/data/

Prerequisites

1. Install Virtualbox: https://www.virtualbox.org/wiki/Downloads

2. Install Vagrant: http://www.vagrantup.com/downloads.html

3. Install necessary Vagrant plugins:

vagrant plugin install vagrant-hostmanager
vagrant plugin install vagrant-cachier

4. Install Ansible

brew install ansible

Getting Started

Start by bringing up the Vagrant machine, it is configured to install everything you need to run the analysis

cd vagrant
vagrant up

Load the sample data by SSHing into the machine and running the tocb.py script.

cd vagrant
vagrant ssh
cd /vagrant
python tocb.py

At the top level of the repo, go into the Spark project, build it, then move the jar file into the Vagrant shared folder:

cd mba
mvn clean package
cp target/mba-1.0-SNAPSHOT.jar ../vagrant/

Once the jar is in place you can SSH into the vagrant machine and run the process:

cd vagrant
vagrant ssh
/opt/spark-1.4.1-bin-hadoop2.6/bin/spark-submit --class com.avalonconsult.mba.MBA /vagrant/mba-1.0-SNAPSHOT.jar

You can access the Couchbase UI at retail.vagrant:8091 with credentials: couchbase//couchbase