Page 96 - Kaleidoscope Academic Conference Proceedings 2022

2022 ITU Kaleidoscope Academic Conference

one of Kadaster’s main organizational aims, namely, to data would have been returned from a range of sources using
provide everyone (from developers and data engineers to a range of schemas, it would have been necessary to
students and interested citizens) with the ability to use and appropriately transform and integrate the data and then
interact with the geospatial data they maintain and publish. render it for display. This, of course, complicates the
An application in this form, specifically, provides a low- development of the application because the application
threshold, gamification-type experience for the user, who requires data transformation and management functionality
only needs access to a smartphone and network connection; rather than simply visualization functionality. Additionally,
arguably improving the accessibility and overall experience the use of SPARQL APIs also means that any addition of
of interacting with Kadaster’s data and the KKG. more information in the future simply requires an updated
query and/or an updated API variable in the application code
7
Currently, the application is made available for download as well as some modification of the way in which the data is
and use within the Netherlands based on the fact that this is rendered. This, therefore, makes the maintenance and
the spatial extent of the data published in the KKG. The extension of the application much easier. A key disadvantage
usability of the application has been tested internally as part in this respect is that if the underlying schema for the source
of the finalization of this innovation project and users are is changed, the queries will automatically break and,
able to provide feedback about the data or application therefore, the application will no longer work as intended.
through the application itself. No formal user experience
research has been carried out to date but is seen as something Finally, and as a consequence of the previously mentioned
that should be considered as a next step in order to improve advantages, this approach also means that the need for
the functionality and data made available through the copying the data for transformation and/or management
application. purposes is negated. As such, data remains at the source and
is always as up to date as possible in the application, again
5. DISCUSSION reducing the complexity in the development and ongoing
maintenance of the application. The only barrier to this
What the previous sections aim to highlight is the novelty, advantage in the approach is the trust that the application
success, and value in implementing an application developer has in the fact that the source has an acceptable
architecture where data storage and application are uptime and is as performant as required for the application
distributed entities. The development and implementation of functionality. A similar trust needs to be had in the fact that
the AR application as described in the previous section the underlying schema used for the data remains stable so
allows for reflection on both the practical advantages and that continuous updates to the query and/or rendering do not
disadvantages of the overall system architecture proposed by need to be done to maintain the application. The authority
[1] (see Figure 3). and dependability of the source, therefore, need to be ensured.

5.1 A standards-based approach 5.2 Application-independent data publication

Firstly, SPARQL, as a standardized SQL-based query From an organizational perspective, the development of the
language, is relatively easy to understand and use, given a AR application at Kadaster also highlights several additional
basic understanding of the SQL language. Data can also be issues worth considering when implementing this approach.
made available through an API in a format that is Firstly, it can be argued that the implementation of this
recognizable to developers unfamiliar with linked data, approach is only appropriate when data is not being
making the data easy to use in a range of applications. The combined in a triple store for a single purpose. Indeed, if data
only disadvantage in this respect is the small learning curve from various, distributed sources is being copied (as a
required to be able to write appropriate, performant subset), transformed into linked data and combined to form
SPARQL queries for integration in the application. a type of Knowledge Graph that only serves one use case or
application because it, for example, only contains data
Secondly, the data required by the AR application is relevant to that use case, this approach is no longer
available through three separate SPARQL APIs by design. appropriate. Doing so would, in essence, be comparable to
This is the key component of interoperability between the more traditional approaches to data storage where data is
KG and application because the APIs used as interfaces are combined and used for a specific purpose. The intention of
open standards-based interfaces. These APIs are then simply the approach, and what this paper hoped to highlight, was
integrated into the application as API variables and the that data from the source could be used in various application
application needs only to render the data for visualization. contexts without the need to perform any further integration
The advantage of this approach is that data is available or transformation to the data itself. Depending on the
through a single data source and using the same schema structure of the organization, it would only be necessary to
making it easy to simply display. Previously, because the have a single KG which various applications use as a data
source where appropriate.

7 https://www.vr-innovations.nl/ar-kadaster-knowledge-graph-
download .

– 50 –