2 examples of how government data linking can work
- By Joab Jackson
- Nov 09, 2009
Although proposals for linking Semantic Web data are all well and good, what can be done with linked data? At the International Semantic Web Conference 2009 in Chantilly, Va., a couple of experts in the art of Linked Data offered a few demonstrations.
In one hands-on tutorial, Dean Allemang, chief scientist at Semantic Web consulting firm TopQuadrant, showed how to use his company's tool, TopBraid Composer, to mash together different government datasets to create entirely new views of data from multiple sources.
The exercise consisted of filtering data downloaded from a government spending Web site, USASpending.gov, by using a set of ontologies describing the structure of the U.S. government as well as the different lines of business as described in the federal enterprise architecture's Business Reference Model (BRM). The ontologies came from the Ontologies for e-Government Web site, at oegov.org.
The goal was to produce a dataset that shows federal IT spending, organized by line of business across each agency.
Allemang first had participants create a dataset at USASpending. Using data drawn from agencies' Exhibit 300 filings, the site lets users create a custom set of data by picking agencies, lines of business, business and types of spending, such as maintenance, upgrades and total spending. USASpending creates a Really Simple Syndication feed with the resulting dataset.
After importing this RSS feed into the TopBraid Composer workspace, Allemang matched areas of spending with associated items in the BRM. These associations could then be modeled in a graph, and they were searchable. One query was: How much was each agency spending for a particular line of business?
One government IT manager in attendance said the demonstration was impressive, noting that such reports usually take a fair amount of work on the part of the Office of Management and Budget and other agencies to create.
Also at the conference, Cory Casanave, chief executive officer of consulting firm ModelDriven.org, demonstrated how government IT architectures can be shared across different formats using the Resource Description Framework as the common element among them. He showed how one model, rendered in the Unified Modeling Language, could be transposed into another format, namely in an Entity Relationship Diagram, using RDF to convey all the important elements.
ModelDriven offers an online demo, called the Enterprise Knowledge Base, that shows how models can be transposed, including those rendered using the FEA Data Reference Model.
Joab Jackson is the senior technology editor for Government Computer News.