To help our clients in choosing a gift we have created an online virtual assistant called Sophie. Sophie stands for "Sophisticated Present Engine".
Sophie is a self-learning system, which, by means of asking simple clickable questions, learns about the person to give a gift to and then presents the top 10 of best fitting gifts. These products can then be bought in our online gift shop. The questions Sophie asks are: man/woman, age, some preference questions (e.g. choose between: shopping or book) and some product-preference questions.
The technology behind Sophie is Collaborative Filtering. This is a proven algorithm to generate recommendations. Amazon.com is known to be one of the first companies to introduce use of this for recommendations. Cadeau.nl reworked the input for this technology to give pre-sale advice to gift shoppers. Sophie is greatly appreciated among the cadeau.nl customers that see it as a valuable help.
How hard can it be to communicate a button click from the browser to the server? Strangely enough, implementing this is still a lot of work in the 21st century. Not with Vaadin, 5 lines of plain Java code and you are done.
As our recommendation engine backend is also written in Java, integration with the Vaadin frontend is just a matter of adding an import statement, no effort at all. The well-written user guide of course helps a lot; see it as a crash course in Vaadin.
"Being able to write our frontend code in plain Java with Vaadin allows us to create complex user interfaces which are cheap to develop, rock solid, maintainable and extendable." - Cadeau.nl CTO Rob van der Veer