This step-by-step tutorial will guide you through the process of building a robust recommendation engine on open technologies. The aim of the tutorial is to get your recommendation engine up and running quickly, and to teach you a bit along the way about powerful technologies for working with data.
The Mortar recommendation engine produces personalized recommendations at scale for companies like MTV, Comedy Central, StubHub, and the Associated Press. It is designed to be:
Open. Every business is different: you know things about your business and your data that can drive better recommendations. This recommendation engine is completely customizable and entirely open source, so you can modify or adapt whatever you need.
Flexible. For this engine, everything is a signal. Every user interaction, content metadata, or even outside streams of data, so you can blend your collaborative filter with your content recommendations with anything you like, and still know why each recommendation was generated.
Portable. You can run this recommendation engine code anywhere. It’s built on widely-adopted open source technologies—Hadoop, Pig, and Python. For this tutorial, we'll show you how to run on the Mortar platform, where all of the components you need to run it in production at scale (deployment, monitoring, visualization, cluster management) are fully handled.
Ready to build a recommendation engine for your business? Let's get started!