A pilot project by the Flanders government to link information on public research projects with dynamic analysis of the research content itself is building for a bright future – one in which academics and businesses will soon be able to access all research in the region at the click of a mouse.
“The basic idea behind the RILOD project evolved in our department a few years ago,” explains Geert Van Grootel, Senior Researcher at the Department of Economy, Science and Innovation (EWI) for the Flemish government. “Why not build an open space where you could search all publicly available research in Flanders – both by headline data and actual research content – and then turn it all into linked, shared intelligence? The faster that information from academic research can move, the shorter the ‘innovation value chain’ – and therefore the faster that we, as a region, can realize a return on investment on our publicly funded research.”
Research Information Linked Open Data (RILOD) is part of a larger government program, known as FRIS (Flanders Research Information Space), designed to integrate information on research conducted at all Flemish universities and one university college. The technical challenges of the RILOD project promised to be huge from the outset. It would require not just the mapping and linking of the structured data already in the public domain (research project information, researchers and institutions from around the region, stored systematically in the Flemish Research Portal) but crawling and analyzing vast quantities of unstructured data too (i.e., the research content itself). “It was an experimental idea at first,” says Van Grootel. “But in a very short space of time, it’s become a reality.”
The resulting “linked open data store” (publicly accessible, albeit still in beta at the moment) provides access to a range of dynamic information on 25,000 projects, 22,000 researchers and 600,000 publications around Flanders. That includes some 200 million “triples” identified within 45,000 of the research texts by IBM’s analytics software, and intelligently correlated (a “triple” being a data entity composed of subject-predicate-object, like “results are promising” or “35 percent chose Option A”).
Crucially, because the product’s IBM analytics can not just read but interpret the text of the research studies, the technology is an intelligent resource. “So a researcher will soon be able to choose terminology, academic theories, people or topics from his research domain and see how they are being used in other research projects; expand out from that to find connections to other associated persons, projects and institutions; and then find intelligent links between key concepts within the research studies and associated publications.”
It should be a big boost to industry as well. “Businesses will be able to locate all this public knowledge without needing to understand how information in academic research is structured – because they’ll be able to use their own terminology, in simple and intelligent search tools, and all in a user-friendly interface.”
And since the underlying data store itself will remain open-source, it could spawn a range of different external applications. “We will build pre-defined visualizations in future iterations, but a private company could easily build their own interface and draw up data from the linked open data store, allowing all this great research to be exploited to its fullest potential,” says Leen Van Campe, Information Architect at EWI.
The complete technology stack has now been test-driven and the project has recently been designated by the Flemish government as a “Flemish Innovation Project” (i.e., an e-government project that drives innovation and facilitates efficiency gains for business and citizens).
The project’s primary metric – generating return on investment through an improved innovation value chain – will take time to be realized fully. But the RILOD will also benefit in the future from other FRIS projects aimed at expanding the region’s information space (which will bring in more stakeholders and more information objects) and at improving the quality on the source side of FRIS. “We honestly believe that, years from now, this is going to have proven transformational,” says Van Grootel. “What we have today is just the beginning.”