NUI Galway – UL Alliance First Annual ENGINEERING AND - ARAN ...

Enabling Federation of Government Metadata Repositories
Gofran Shukair 1 , Nikos Loutas 1 , Fadi Maali 1 , Vassilios Peristeras 2
1 Digital Enterprise Research Institute, NUI Galway
(firstname.lastname)@deri.org
2 EC, DG for Informatics, Interoperability Solutions for European Public Administrations
Abstract
Data models, taxonomies, ontologies, code lists and
semantic data exchange formats are the key resources
for creating interoperable data exchange. These
resources exists in several national repositories that
differ in scope and in target group they are addressing
and they are implemented using different technologies
and expose different user interfaces to the end user.
However the semantic content they include can often be
reused even bypass the domain they were originally
designed for. Seamless data exchange has been the
motivation to enable interoperability of these different
repositories in order to reach a federation of
repositories where users can search for resources or
assets in all the available repositories using a unified
user interface.
1. Introduction
Resources, such as data models and schemata,
taxonomies, ontologies, code lists are the means for
seamless data exchange. In the context of this work, we
use the term asset to refer to these types of resources.
When it comes to government data, assets exist in
isolated national e-government metadata repositories
that differ (i) in scope, (ii) in target group they are
addressing, (iii) are implemented using different
technologies and (iv) expose different user interfaces to
the end user. The semantic content they include can
often be reused even bypass the domain they were
originally designed for. But the physical isolation of
these repositories and the heterogeneity of the assets
hamper the reusability of common concepts.
Hence, this research aims to (a) define a common
data model to describe the assets of national egovernment
metadata repositories and (b) to develop a
proof-of-concept prototype showcasing the added-value
of federation through cross-repository asset discovery
and retrieval using a single point of access.
2. Common Model Development
This section introduces the Asset Description
Metadata Schema (ADMS) and discusses its
implementation using linked data and Semantic Web
technologies. Thus, ADMS is implemented as an RDF
Schema vocabulary which includes the core elements
and the attributes needed to model an asset. We reuse
properties from existing linked data vocabularies (e.g.
DCAT 1 , DCMI 2 ) in the definition of an asset (Figure 1).
1 F. Maali, R. Cyganiak, V. Peristeras. Enabling interoperability of
government data catalogues. In Electronic Government, LNCS, 2010.
2 http://dublincore.org/
113
An asset is a container dedicated to group artefact
types that enable cross-repository interoperability and
seamless data exchange. It is defined as a subclass
dcat:Dataset with additional properties like dc:publisher
and dc:spatial. The property adms:release is used for
differentiating between the versions of each asset and
the asset itself, Each asset will be related to its latest
release using the property adms:release. Each release
will be related to its previous one using. dc:hasVersion.
Figure 1. ADMS Model in RDF
3. Proof of concept Architecture
Figure 2 shows the proposed architecture where
assets from different national e-government metadata
repositories can be queried and retrieved through a
single point of access (SPARQL endpoint). Valueadded
services, such as data mashups and faceted
browsers, and third party applications can be built on
top of this architecture. ADMS defines the
administrative and descriptive metadata of the assets
housed in the repositories. These assets can then
queried and retrieved using querying API. Repositories
can be connected to the federation and publish their
assets’ metadata using the Publishing API and services.
Figure 2. Proof of Concept prototype
Acknowledgments. This work is a joint effort between
]init[, EC ISA, NUI Galway and the University of
Macedonia. It is funded in part by Science Foundation
Ireland under Grant No. SFI/08/CE/I1380 (Lion-2).