The University of Southampton is at the forefront of work in the semantic web and has produced various open and linked data sites and tools over the years (Blog). Now it has embarked on a project to make large amounts of the organisation's own data available as Linked Open Data, in a sustainable and useful way. Many parts of the University are involved including research, teaching, catering, finance, estates and transport. This has then been used to build a number of useful applications for the University staff, students and visitors. A very pragmatic approach has been taken, and rather than just publish the RDF, a website has been published over the top of the data which creates immediate value from data which was previously buried in spreadsheets. The focus has been on creating a agile, manageable and sustainable system rather than something perfect but brittle. The data comes from a mix of sources; various databases, 4 different EPrints repositories and a large number of google-docs spreadsheets, each maintained by the appropriate data owner (Example).
All the tools used to build the service are free open source software. We use a virutal machine running Ubuntu Linux, store our data in 4-store. The smaller datasets are edited using Google Spreadsheets. We prepare our data using OpenOrgGrinder (created by us but open source) which converts spreadsheets into RDF, XSLT and rapper. To provide useful views of our data we use ARC2 PHP library, Graphite PHP Library (created by us) and Google Maps. The public SPARQL endpoint sits in front of the real endpoint and adds extra features not available from 4store.
