Product Search using Linked Data

Kiln October 10, 2018 15:00 - 15:15

Bookmark and Share

John Walker and Herman Elenbaas.  

Abstract:  For semiconductor manufacturers such as Nexperia it is important to be able to answer information requests from customers in a timely manner. With a portfolio of over ten thousand product models, being able to quickly and easily find the answers to frequently-asked questions can be the difference: efficiency wins. In this paper we explain how Linked Data technology was used to integrate product data from multiple heterogeneous data sources using the Dydra platform to enable rapid declarative provisioning of APIs. We also explain how an ETL process was created to populate a search index using SPARQL and JSON-LD Framing to generate JSON-LD documents that are indexed in Elasticsearch to enable full-text and faceted search. A Product Search application was developed with focus on UX/UI resulting in a fast and responsive user experience. The application has been rolled out globally across the organization with user in various disciplines from sales and marketing to quality and research.

Keywords:  RDF;  SPARQL;  JSON-LD;  ETL;  product search