=Paper=
{{Paper
|id=Vol-3632/ISWC2023_paper_423
|storemode=property
|title=GeoChangeViz: Visualizing Knowledge Graphs about Changes in Geographical Divisions
|pdfUrl=https://ceur-ws.org/Vol-3632/ISWC2023_paper_423.pdf
|volume=Vol-3632
|authors=Camille Bernard,Matthieu Viry,Marlene Villanova,Jerome Gensel
|dblpUrl=https://dblp.org/rec/conf/semweb/BernardVVG23
}}
==GeoChangeViz: Visualizing Knowledge Graphs about Changes in Geographical Divisions==
https://ceur-ws.org/Vol-3632/ISWC2023_paper_423.pdf
GeoChangeViz: Visualizing Knowledge Graphs about
Changes in Geographical Divisions
Camille Bernard1,∗,† , Matthieu Viry2,† , Marlène Villanova1,† and Jérôme Gensel1,†
1
Univ. Grenoble Alpes, CNRS, Grenoble INP, LIG, Grenoble, France
2
Université de Paris, UAR RIATE, CNRS, F-75013, Paris, France
Abstract
Geographical divisions (GD) being administrative, parcel or census divisions frequently change over
time. When such changes are not well documented, they introduce biases in the understanding of the
evolution of such a given area. To cope with this problem, we have designed a Knowledge Graph (KG)
that allows for the description of many GDs in the world and the changes they undergone over time. In
this demo, we present a Web application called GeoChange-Viz that connects to this KG and provides
spatial planners with an innovative tool for geovisualizing and understanding territorial changes over
time.
Keywords
Spatiotemporal Knowledge Graph, Knowledge Graph visualization, Linked Geospatial Data
1. Introduction
All over the world, different kinds of Geographical divisions (GD) (being administrative, electoral,
census or parcel divisions) are defined by (local) authorities. Following administrative or political
reforms and decrees, the boundaries, name or identifier of these GD change over time. For
instance, in the United States, census units are frequently redefined to maintain an equivalent
number of inhabitants per unit. When such changes are not well documented, they are major
obstacles to the understanding of a study area. As a result, decision-makers and spatial planners
need tools to monitor these territorial changes and avoid many errors when analyzing data
series collected on divisions that have changed. In this demo, we present the GeoChange-Viz
Web application designed to assist spatial planners in the monitoring of changes their study
areas have undergone. GeoChange-Viz is based on a Knowledge Graph (KG) we have designed
[1], called GeoChange KG1 . This GeoChange KG describes GD and their evolution over time. It
ISWC 2023 Posters and Demos: 22nd International Semantic Web Conference, November 6–10, 2023, Athens, Greece
∗
Corresponding author.
†
These authors contributed equally.
Envelope-Open camille.bernard@univ-grenoble-alpes.fr (C. Bernard); matthieu.viry@cnrs.fr (M. Viry);
marlene.villanova@univ-grenoble-alpes.fr (M. Villanova); jerome.gensel@univ-grenoble-alpes.fr (J. Gensel)
GLOBE https://lig-membres.imag.fr/bernardc/ (C. Bernard); https://mthh.github.io/portfolio/ (M. Viry);
https://lig-membres.imag.fr/villanov/ (M. Villanova); https://lig-membres.imag.fr/gensel/ (J. Gensel)
Orcid 0000-0003-2246-6568 (C. Bernard); 0000-0002-0693-8556 (M. Viry); 0000-0002-7242-6102 (M. Villanova);
0000-0003-1398-7118 (J. Gensel)
© 2023 Copyright c 2023 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
CEUR Workshop Proceedings (CEUR-WS.org)
Proceedings
http://ceur-ws.org
ISSN 1613-0073
1
http://steamerlod.imag.fr/sparql
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
provides rich descriptions of GD changes, using terms such as Split, Merge, Name or Identifier
change, etc. Moreover, the GeoChange KG is enriched with data from the Open Data Web. The
GeoChange-Viz application exploits the structure of this KG to show, by means of innovative
graphical components, GD changes along the lifelines of Territorial Units (TU) (i.e., each division
of the geographical space in the GD).
2. The GeoChange Knowledge Graph
In [2], we have presented the TSN2 (Territorial Statistical Nomenclature) ontology whose terms
are generic enough to allow for the description of any multi-levels GD that can be found
around in the world. This ontology goes hand in hand with the TSN-Change ontology 3 which
consists of a generic vocabulary for the description of changes in GD (using terms such as
name/identifier/boundaries change, merging or splitting of TUs, etc.). In this vocabulary,
distinctions are made between changes where the identifiers of the geographical areas persist
after change or not (e.g, fusion vs integration). This ensure a clear understanding of the changes
and avoid misinterpreting data when a TU does not change its identifier over the versions of
a GD, even though it has undergone major geographical changes. The TSN and TSN-Change
ontologies are based on (standards) ontologies and frameworks such as GeoSPARQL [3], OWL-
Time [4], PAV [5], DCTerms4 , BFO [6]. In [7], we have presented a framework named Theseus
to populate the TSN and TSN-Change ontologies. It is based on an algorithm defined for the
detection and semantic annotation of territorial changes. Using the Theseus framework, the
GeoChange KG can be populated with the description of various GD and versions. So far, the
graph contains the description of diverse GD (covering respectively the evolution, over a given
period, of GD in Switzerland, Australia, European Union and France). This graph, dedicated to
the description of changes in GD, follows a specific structure that is named ”Multi-levels change
graph”. This structure is inspired by [8] while adding the notion of ”multi-level”, since changes
that occur at one given level of the GD (e.g., state level) propagate to changes impacting lower
divisions (e.g., districts). The scenario of the demo presented here is based on a French GD
called AdminExpress 5 , in versions 2015 and 2016, which is composed of 3 French administrative
levels: regions, departments and municipalities6 . The Figure 1 is an extract of the GeoChange
KG showing a change event that has occurred at level 2 (municipalities, see label L2 of the
nodes in Figure 1) in this GD, between years 2015 and 2016: the integration of a municipality
(circled green node with label V2015_L2_38021) into another (circled green node with label
V2015_L2_38225) municipality, results in the new municipality with code V2016_L2_38225. In
the following section, we show how the GeoChange-Viz application can be used to geo-visualize
this graph, and make it available to an audience with no technical background in KG.
2
http://purl.org/net/tsn#
3
http://purl.org/net/tsnchange#
4
https://www.dublincore.org/specifications/dublin-core/dcmi-terms/
5
Source https://geoservices.ign.fr/adminexpress
6
Please, note that in the demo video, only an extract of this GD (covering the French region AURA) is loaded from
the GeoChange KG, under the name GeoflAURA.
3. The GeoChange-Viz Web Application
The GeoChange-Viz application connects to the GeoChange KG and guides users, step by step,
in their understanding of territorial changes. In a similar approach to the SexTant application
[9], GeoChange-Viz focuses on the visualisation of the temporal dimension of linked geospatial
data. In [10], the authors present a generic model for delivering time-evolving 3D city models for
Web visualization. They describe transitions between two versions of a geographic feature. They
propose to link documents to these transitions which help to give elements of proofs regarding
how geographic features evolve in time. Our approach is also to describe the changes (or
transitions) between two versions of a GD, by semantizing these changes and searching the Web
for documents that can explain their causes. Indeed, each year, Urban Planning Authoritative
Services (UPAS) must identify where, when, how much, why and how territorial changes took
place, in order to maintain their expertise on their study area and to transfer their statistical
data into the latest version of the GD. In particular, knowing how the TUs change over time
(i.e., the nature of the change event) helps them in determining the appropriated statistical data
processing for this transfer. In this demo, we will progressively show how such an UPAS can
answer these where, when, how much, why and how questions from the application.
Figure 1: Excerpt of the GeoChange KG showing the merger of two municipalities between 2015 and
2016, in the french AdminExpress GD (bottom part of the graph: integration of V2015_L2_38021 into
V2015_L2_38225 resulting in V2016_L2_38225)
Where? When? How much? Within GeoChange-Viz, several UIs have been designed to
understand territorial changes step by step. In particular, the UI named Synchronised maps
helps the user to quickly identify where the changes have taken place, using a set of colors
highlighting the changes on the map according to their nature (on the contrary, unchanged
units are kept grey) (see Figure 2). The two synchronized maps are dated (giving information
Figure 2: Visualization of the KG shown by Figure 1 in the GeoChange-Viz App: the synchronized
maps and the Change Bridge component render the merger of two municipalities in France, between
2015 and 2016.
on the When): the map on the right shows the TUs before the change event and, on the left, the
TUs after the change event. Below the map, a bar graph quantifies the number of TUs impacted.
Also, the Change Catalog tab lists and quantifies the change events, according to their nature
(providing information on the How much).
Why? How? By selecting a TU on the map, the user accesses its description, enriched with
external descriptions (e.g., links to LOD Wikipedia, Geonames and Wikidata representations of
the TU) that, most of the time, confirm the territorial change and explain it. Similarly, when
the information is available on the Web, the GeoChange KG connects to legislation (providing
information on the Why). Also, a UI component named Change Bridge gives the user access to a
visualization of a bunch of the GeoChange KG with semantic labels describing the nature of the
change event (providing information on the How). For instance, as shown in the Figure 2, the
user has selected a municipality on the right map and, in the Change Bridge component, s-he
learns that this TU has merged with another. As one of the merging TUs keeps its identifier after
the change event, the change event is described as an Integration in the GeoChange KG. This
precision is important as it may alert users and avoid misinterpretation of statistical data that
refer to a TU using its unchanged identifier while its boundaries and, consequently, probably
its number of inhabitants, have changed.
4. Conclusion
In this demo, we have shown how the use of semantic Web technologies with UI can help spatial
planners in understanding how GD evolve over time. KGs are here first used to represent GDs
and the changes they are subject to, and then enriched by contextual information that brings
explanations for the observed evolution. Our future work will consist in carrying out a study
on the use of the application in order to reinforce the key functionalities of the application.
Demo link: https://videos.univ-grenoble-alpes.fr/video/26646-demo-GeoChange-Viz-iswc-2023mp4
Password: 2023-geochange-iswc
Acknowledgments
This work is supported by LINKSIUM, SATT Grenoble Alpes.
References
[1] C. Bernard, Immersing evolving geographic divisions in the semantic Web, Theses, Uni-
versité Grenoble Alpes, 2019. URL: https://theses.hal.science/tel-02524361.
[2] C. Bernard, M. Villanova-Oliver, J. Gensel, H. Dao, Modeling changes in territorial partitions
over time: Ontologies tsn and tsn-change, in: Proceedings of the 33rd Annual ACM
Symposium on Applied Computing, SAC ’18, ACM, 2018, pp. 866–875. URL: http://doi.acm.
org/10.1145/3167132.3167227. doi:1 0 . 1 1 4 5 / 3 1 6 7 1 3 2 . 3 1 6 7 2 2 7 .
[3] M. Perry, J. Herring, OGC GeoSPARQL - A Geographic Query Language for RDF Data
(2012) 75. URL: https://www.ogc.org/standard/geosparql/.
[4] S. Cox, C. Little, Time Ontology in OWL - W3C Recommendation 19 October 2017, 2017.
URL: https://www.w3.org/TR/owl-time/.
[5] P. Ciccarese, S. Soiland-Reyes, K. Belhajjame, A. J. Gray, C. Goble, T. Clark, Pav ontology:
provenance, authoring and versioning, Journal of biomedical semantics 4 (2013) 1–22.
doi:1 0 . 1 1 8 6 / 2 0 4 1 - 1 4 8 0 - 4 - 3 7 .
[6] P. Grenon, B. Smith, SNAP and SPAN: Towards Dynamic Spatial Ontology, Spatial
Cognition & Computation 4 (2004) 69–104. doi:1 0 . 1 2 0 7 / s 1 5 4 2 7 6 3 3 s c c 0 4 0 1 _ 5 .
[7] C. Bernard, M. Villanova-Oliver, J. Gensel, Theseus: A framework for managing knowledge
graphs about geographical divisions and their evolution, Transactions in GIS 26 (2022)
3202–3224. doi:1 0 . 1 1 1 1 / t g i s . 1 2 9 8 8 .
[8] T. Kauppinen, E. Hyvönen, Modeling and reasoning about changes in ontology time series,
in: Ontologies, Springer, 2007, pp. 319–338. doi:1 0 . 1 0 0 7 / 9 7 8 - 0 - 3 8 7 - 3 7 0 2 2 - 4 _ 1 1 .
[9] K. Bereta, C. Nikolaou, M. Karpathiotakis, K. Kyzirakos, M. Koubarakis, Sextant: Visualizing
time-evolving linked geospatial data, in: E. Blomqvist, T. Groza (Eds.), Proceedings of
the ISWC 2013 Posters & Demonstrations Track, Sydney, Australia, October 23, 2013,
volume 1035 of CEUR Workshop Proceedings, CEUR-WS.org, 2013, pp. 177–180. URL: https:
//ceur-ws.org/Vol-1035/iswc2013_demo_45.pdf.
[10] V. Jaillot, V. Rigolle, S. Servigne, J. Samuel, G. Gesquière, Integrating multimedia docu-
ments and time-evolving 3d city models for web visualization and navigation 25 (2021)
1419–1438. URL: https://onlinelibrary-wiley-com.sid2nomade-1.grenet.fr/doi/10.1111/tgis.
12734. doi:1 0 . 1 1 1 1 / t g i s . 1 2 7 3 4 , publisher: John Wiley & Sons, Ltd.