Datasentiment have created an open source demo of their recommendation engine. There’s a full rundown of how to get it working, what data to use and suggestions of further development that could be done.
The demo on Github works on a system of 5000 customers, 45000 items and a million transactions.
The idea started life from a blog post I did on the importance of recommendations in retail for 2013. The revolution will be personalised and recommended by someone else.
As ever I’m always happy to listen to your thoughts and feedback, you can email me at firstname.lastname@example.org or on twitter @datasentiment/@jasebell