The results of the Semantic Web Developer Challenge, co-sponsored by XSB and SemanticWeb.com and launched at this summer’s Semantic Technology and Business Conference, are in: The $5,000 prizewinner was a team of two, Greg Varga and Siraj Bawa, from Vanderbilt University. There were two runners-up: One was a team from Stony Brook University, comprised of Mrinal.Priyadarshi, Anurag Choudhary, and Paul Fodor, and the other was Roman Sova from consulting firm Good Monster.
The aim of the Challenge was to build sourcing and product life cycle management applications leveraging XSB’s PartLink Data Model, which the company developed as a project for the Department of Defense Rapid Innovation Fund. The model uses semantic web technology to create a coherent Linked Data model for all part information in the Department of Defense’s supply chain – which includes about 40 million component parts, their manufacturers and suppliers, materials, technical characteristics and more.
The large collection of engineering product information has potential beyond DoD use alone. “The current size of the Part Link triple store is well over a billion triples — maybe 1.3 billion,” says Alberto Cassola, vp sales and marketing at XSB. “For the industrial sector it may very well be one of the largest efforts of its kind.”
The six judges of the challenge included Dennis Wisnosky ,former DOD CTO for Business and Chief Architect, and currently a Semantic Web consultant as well as two members of the Department of Defense Logistics Agency.
The winning team from Vanderbilt University individually brought computer science and economics savvy to their entry, using the Producer Price Index to bring all suppliers’ historical prices up to today’s dollars. Their app normalized prices for each part number listed in past purchase orders and then rated each supplier found for that part according to the computed pricing index. They also used turnaround delivery times to come up with another index of how good one supplier is in that respect vs. competitors, according to XSB CTO Chris Rued. “They came up with a composite index combining those two to allow the user to decide which they care about more” – price or speed – for their procurement decisions, he says.
“The visual interface was really neat because of the slider that provided an animated graphical representation” to surface and make actionable the data that is most important to the end user, says Cassola.
The Stony Brook app enables users to punch in the part number or manufacturer serial number to return lists of all the suppliers that offer a particular part with particular attributes and their addresses; lets users view locations as pins via Google Maps; and calculates the distance between that location and delivery points. Users also can see each part’s current prices. Part Link contains many standardized attributes, and “the Stony Brook team is the only one that really leveraged the ontology in that way to create an app that would not have been possible otherwise,” says Rued. “They used the item characteristics, the properties of the items, described in the ontology to build the app.”
While these first two apps were mostly focused on the user interface, Rued says, the Good Monster entry tapped into the API angle. It worked from the medical readiness point of view, so that it would be possible for users at agencies such as the U.S. Department of the Navy to integrate with their surrounding systems to easily satisfy requirements pointed out by audits in that area for healthcare items, such as first aid kits.
Sova “built the API where someone could load up items of interest [that the audit identified were lacking] with descriptions and parts numbers, and search to see who could supply the items, and what is their location,” Rued notes. “It hooks to the U.S. Postal Service API to calculate delivery estimates from supplier locale to destination. The idea is setting up the API to let existing systems around it integrate with it and answer questions about real problems.”
Next steps for XSB will include identifying additional commercial opportunities around Part Link, such as providing it as a platform-as-a-service to connect data across an enterprise, its suppliers and customers. “Beginning to develop apps was an early demonstration of the ease of using this data platform to create useful apps for logistics and those managing large supply chains,” he says. “So this is really to demonstrate the long-term vision of building the industrial semantic web, and 2015 is the year to move into practicable applications of that which can be used in the Department of Defense and commercial arenas.”