Page 92 - Kaleidoscope Academic Conference Proceedings 2022

2022 ITU Kaleidoscope Academic Conference

understood and maintained and offers the possibility of
adding more data options over time in a relatively flexible
manner. From the organizational perspective, the KG
improves the governance of the data contained within. From
the perspective of the end user, the application offers a low-
threshold, user-friendly manner in which to access, interact
and provide feedback on the geospatial data published by
Kadaster. Overall, this paper will argue that applications
such as the AR application exemplify the opportunities that
such an architectural approach holds for a decoupled,
flexible ecosystem and how such an application will act as
key drivers for further investment in implementing this
approach more widely; particularly within the context of
governmental organization. Figure 1 – Triple structure (subject-predicate-object)
using a geospatial example
2. KNOWLEDGE GRAPHS FOR DATA
INTEROPERABILITY The semantic web, in facilitating the creation of linked,
machine-readable data on the web, makes use of RDF, the
The Knowledge Graph (KG), a data representation model in associated schema (RDFS) and a range of Web Ontology
which data is stored as a graph comprised of nodes and edges Languages (OWLs) as the core web technologies which
mapped to ontologies to form a semantic network, was structure data on the web. These web technologies formalize
introduced as a means of connecting and integrating data the semantics of data from various domains using defined
from different sources. Indeed, a knowledge graph or ontologies and capture these semantics in triples using
semantic network can be defined as representing ‘a network Unique Resource Identifiers (URIs) for nodes and relations
of real-world entities – i.e. objects, events, situations or [4, 5], enabling location-independent cross-machine
concepts – and illustrates the relationship between them’ [2]. readability and interaction [6]. A single triple can be
Although the concept of the knowledge graphs is not recent, connected iteratively with other triples, forming an expanded
the term was first popularized by Google in 2012 and has graph data model known as the knowledge graph. These
since formed part of technological solutions from various connected triples can be semantically enriched and placed in
multinational corporations and as part of search engines such context through the application of (domain) ontologies and
as Google and Yahoo. During implementation, a knowledge connection to other knowledge bases using semantic web
graph makes use of various data management models technologies and standards.
including the traditional database model, a graph model and
a knowledge base model wherein the formal semantics (from The interoperability of data made available on the web, such
various domains) are defined [3]. as data made available as part of a knowledge graph, is
supported by the efforts of the World Wide Web Consortium
In general, the knowledge graph makes use of the Resource (W3C), the main standardization organization for the
Description Framework (RDF) structure for data worldwide web. This consortium publishes and maintains a
representation and, as such, there are three main components variety of open standards, including common linked data
which make up said graph, namely nodes, edges and labels. standards such as OWL, RDF, SPARQL, PROV and SKOS,
Objects, places or persons, for example, can be represented which aim to foster compatibility and agreement in the
using a node and the relationships between these nodes are publication of data on the web. Where reusability of open
represented by an edge. This node-edge-node structure, or standards is high, there is a shared understanding of the
the subject-predicate-object structure, is known as a triple, a meaning of data published using these standards and,
data representation structure which underpins the publication therefore, interoperability and potential for reuse of datasets
of data as linked data. An example of this triple structure in available on the web is higher.
the context of geospatial data is provided in Figure 1.
The following section discusses the architecture used in the
creation and ongoing development of the Kadaster
Knowledge Graph (KKG). The use of open standards
published by the W3C in the modeling of the KKG is a
central element of this architecture. The standards used in
this solution include OWL-based ontologies, RDF and
RDFS for data representation as well as elements of the
SKOS and PROV vocabularies for knowledge organization
and provenance links between source datasets and the KKG.
Naturally, the resulting linked data is also made available
through a SPARQL endpoint, another example of a
standards-based implementation within the solution
architecture.

– 46 –