Database Reference
In-Depth Information
8
Using Linked Data
8.1 INTRODUCTION
In Chapter 7, we covered the process of creating and publishing Linked Data from
Merea Maps' own Geographical Information (GI). Now we move on to cover a
number of issues surrounding the use of Linked Data: how to query, interlink, and
create business value from it. We start by exploring the reasons why an organi-
zation with GI should consider investment in Linked Data publishing and discuss
some Linked Data business models. Section 8.3 explains the RDF query language
SPARQL, which can be used to construct queries into a triple store or published
dataset via a SPARQL endpoint. The bulk of the chapter discusses the process of
selecting links from Merea Maps' own dataset to third parties, including quality
versus volume trade-offs between manually curated and automatically generated
links and specifying the context in which a link might be valid. When we look
at data integration through linking, it becomes apparent that the RDFS (Resource
Description Framework Schema) language is limited in its ability to fully express
the knowledge required to really know whether and when two things are the same
or what other relationship exists between them. Motivated by these shortcomings,
Chapter 9 then moves on to introduce methods of encoding more nuanced informa-
tion in ontologies using the OWL Web Ontology Language.
8.2 BUSINESS MODELS FOR LINKED DATA
So, why should a GI data publisher go to the expense and effort of expressing its GI
as Linked Data? While the Linked Open Data movement encourages the publication
of data with an open license, for publishers who have spent years surveying and col-
lecting data and whose business relies on selling data products, this is not always a
realistic option. There are a number of ways to generate value from Linked Data, as
explained by Scott Brinker
1
and Leigh Dodds.
2
8.2.1 S
ubSidy
M
odelS
The subsidy model is the business model followed by, among others, the U.K.
government's Linked Data initiative, the BBC, and the U.S. Census bureau. In this
scenario, the publisher, usually a public sector, educational, or charitable organiza-
tion, is funded (sometimes by donated time, such as in the case of GeoNames) to
produce the Linked Data for public benefit.
137
