DG Joint Research Center
13th of March 2014
Study and guidelines on Geospatial Linked Data as part of ISA Action 1.17Resource Description Framework
Danny VandenbrouckeDiederik Tirry
Click to edit Master title style
Agenda
Introduction
Context
3
2
2
Literature study results
4
5
Sharing your experience – Issues/Challenges?
1
Next steps - Experiments
Click to edit Master title style
Introduction
3
We will provide an outline of the study and our work so far. This webinar is also an opportunity to provide feedback and to exchange experiences: where could INSPIRE RDF be used in e-government?
METHODOLOGIES
WEBINAR II Guidelines on methodologies
Presentation of the first version of the guidelines for a common RDF Vocabulary for INSPIRE data and an approach to PID governance: opportunities for community-led improvements and next steps towards an official INSPIRE encoding.
Today’s webinar
1. Setting the scene
2. Developing proposals
3. Refinement & recommendations
Resource Description Framework
(RDF)
Click to edit Master title style
Agenda
Introduction
Context
3
2
4
Literature study results
4
5
Sharing your experience – Issues/Challenges?
1
Next steps - Experiments
Click to edit Master title style
Context
5
This study has been prepared in the context of the Interoperability for European Public Administrations (ISA) Programme and, in particular A Reusable INSPIRE Reference Platform (ARE3NA, ISA Action 1.17)
Connecting Geospatial DataThis study should provide:
1. Shared evidence about the current status in Europe of linked (geospatial) data related to INSPIRE.
2. An initial common/agreed methodology and guidelines towards RDF encodings for INSPIRE
3. Recommendations for how location PIDs could be governed for INSPIRE and other relevant activities.
Click to edit Master title style
Context
12/12
6
INSPIREInteroperability of geospatial data sets and services through harmonised data models and encodings for the exchange of data related to one of the 34 spatial data themes
1. Data models using UML at conceptual level2. Encoding using GML based on encoding rules
Several European project and national initiatives using publishing geospatial data as Linked Data
Using the Resource Description Framework (RDF)
However No agreed rules or guidelines on how to create
such RDF vocabularies from the UML models
Click to edit Master title style
Agenda
Introduction
Context
3
2
7
Literature study results
4
5
Sharing your experience – Issues/Challenges?
1
Next steps - Experiments
Click to edit Master title style
Current State and State of the Art
8
Literature selection and review approach
Compile and examine relevant, existing studies • External publications from presentations, technical reports, papers and books
(see list)
Describe different projects/initiatives at the European, national and sub-national levels, in the form of an overview table, including the approach followed • Most of the literature is related to (pilot) projects that took/are taking place in
some Member States, among others The Netherlands, UK, Germany, Spain, Italy and Belgium or in European projects such as GeoKnow
A selection of different projects/initiatives will be analyzed in more detail based upon agreed criteria • The focus for this part of the analysis will be on the technical aspects of the
transformation of UML models into RDF
The literature analysis is currently ongoing• The literature overview for RDF/UML is presented in the table below
Click to edit Master title style
Literature study results
9
Author(s) Date Title Type Description
Folmer, E., Reuvers, M., & Wilko, Q. 2013 Pilot Linked Open Data Nederland (NL) B Source: http://www.pilod.nl/doc/boek2.pdf Hart, G., & Dolbear, C. 2013 Linked Data: a Geographic Perspective B Source: Boca Raton: CRC Press, Taylor & Francis GroupHeath, T., Hausenblas, M., Bizer, C., Cyganiak, R., & Hartig, O.
2008How to Publish Linked Data on the Web W
Source: http://events.linkeddata.org/iswc2008tutorial/
Jentzsch, A. 2011 LOD Cloud Diagram as of September 2011 W Source: http://en.wikipedia.org/wiki/File:LOD_Cloud_Diagram_as_of_September_2011.png
W3C 2013 Linking Open Data - W3C SWEO Community Project. W Source: http://www.w3.org/wiki/SweoIG/TaskForces/CommunityProjects/LinkingOpenData
Folmer, E. 2013 Introductie tweede Linked Open Data Pilot PPT Source: http://www.geonovum.nl/sites/default/files/Presentatie%200%20ErwinFolmerv4.pdf
Farazi, F. et al. 2012Trentino government linked open geodata: first results PPT Source:
http://inspire.jrc.ec.europa.eu/events/conferences/inspire_2012/presentations/14.pdf Lopez-Pellicer, F.J., Florczyk,
A.J., Nogueras-Iso, J., Muro-Medrano, P.R. & Zarazaga-Soria, F.J.
2011Linked Open Data for INSPIRE: From 3 to 5 star geospatial data
PPT Source: http://inspire.jrc.ec.europa.eu/events/conferences/inspire_2011/presentations/170.pdf
Schade, S. & Lutz, M. 2010 Opportunities and Challenges for using Linked Data in INSPIRE
P Source: http://publications.jrc.ec.europa.eu/repository/handle/111111111/15247 Tschirner, S., Scherp, A. & Staab, S. 2011 Semantic access to INSPIRE: How to publish and query
advanced GML dataP Source: http://ceur-ws.org/Vol-798/paper7.pdf
Vanbockryck, J. & Robbrecht, J. 2012 Start to Link, a practical POI approach PPT Source: http://www.poweredbyinspire.eu/documents/0503-linkeddata-robbrecht.pdf Beckers, V. & Tirry, D. 2013 Linked Open Data: Pilot Project of NGI-Belgium R Source: currently internal (report waiting for approval)De Keyzer, M., Loutas, N., Colas, C. & Goedertier, S. 2013
TM1.2. Introduction to Linked Data (en) PPT Source: https://joinup.ec.europa.eu/community/ods/document/tm12-introduction-linked-data-en
De Keyzer, M., Loutas, N. & Goedertier, S. 2013
TM1.3. Introduction to RDF & SPARQL (en) PPT Source: https://joinup.ec.europa.eu/community/ods/document/tm13-introduction-rdf-sparql-en
Loutas, N., De Keyzer, M. & Goedertier, S. 2013
TM2.3. Design & Manage Persistent URIs (en) PPT Source: https://joinup.ec.europa.eu/community/ods/document/tm23-design-manage-persistent-uris-en
van den Brink, L., Janssen, P., Quak, W. 2013
From geodata to linked data: Automated Transformation from GML to RDF
BC Source:
http://www.pilod.nl/wiki/Boek/BrinkEtAl-GML2RDF
Click to edit Master title style
Literature study results
10
Author(s) Date Title Type Description
Archer, P., Loutas, N. & Goedertier, S. 2013
Cookbook for translating Data Models to RDF Schemas
R Source: https://joinup.ec.europa.eu/community/semic/document/cookbook-translating-data-models-rdf-schemas Colas, C., Goedertier, S.,
Kourtidis, S, Loutas, N. & Rubino, F.
2013Core Location Pilot - Interconnecting Belgian Address Data
R Source: https://joinup.ec.europa.eu/asset/core_location/document/core-location-pilot-interconnecting-belgian-address-data
Athanasiou, S. et al. 2013Deliverable 2.2.1 Integration of External Geospatial Databases
R Source: http://svn.aksw.org/projects/GeoKnow/Public/D2.2.1_Integration_of_Geospatial_Databases.pdf
Williams, H. et al. 2013Deliverable 2.3.1 Prototype of built in Geospatial Capabilities
R Source: http://svn.aksw.org/projects/GeoKnow/Public/D2.3.1_Prototype_of_Built-in_Geospatial_Capabilities.pdf
Ngonga, A., Sherif, M. & Hassan, M. 2013
Deliverable 3.1.1 Development of First Prototype for Spatially Interlinking Data Sets
R Source: http://svn.aksw.org/projects/GeoKnow/Public/D3.1.1.Development_of_First_Prototype_for_Spatially.pdf
Wauer, M., Both, A., Stadtler, C. & Isele, R. 2013
Deliverable 6.1.2 Report on Customer Data Preparation
and Transformation for Linked Data Usage
R Source: http://svn.aksw.org/projects/GeoKnow/Public/D6.1.2_Customer_data_preparation.pdf GeoKnow 2013 Task 2.7: Exposing INSPIRE data as Linked Data. R Source: http://geoknow.eu/t2-7.html
OGC 2012 GeoSPARQL - A Geographic Query Language for RDF Data
S Source: http://www.opengeospatial.org/standards/geosparql
Abbas, S. & Ojo, A. 2013Towards a Linked Geospatial Data Infrastructure. P Source: EGOVIS/EDEM 2013: 196-210. DOI: 10.1007/978-3-642-40160-
2_16Kerry Taylor, K., Lefort, L., Squire, G., Walker, G., Woolf, A., Shu, Y., Ratcliffe, D., Cox, S., Haller, A.
2014
Developing Ontologies for Linked Geospatial Data P
http://www.w3.org/2014/03/lgd/papers/lgd14_submission_41.pdf
Tsinaraki, C., Stavrakantonakis, I. & Christodoulakis, S.
2007XS2OWL: Representation of XML Schemas in OWL syntax
Whttp://www.music.tuc.gr/projects/sw/xs2owl/
Click to edit Master title styleINSPIRE community• Tschirner, S., Scherp, A. & Staab, S. -
Semantic access to INSPIRE: How to publish and query advanced GML data
• van den Brink, L., Janssen, P. & Quak, W. - Linking spatial data: automated conversion of geo-information models and GML data to RDF
• Hobona, G., Brackin, R. – OGC OWS-8 Cross Community Interoperability (CCI) Semantic Mediation Engineering Report
• ISO/TC 211 – DIS-19150-2 - Ontology - Part 2: Rules for developing ontologies in the Web Ontology Language (OWL)
• GeoKnow - Task 2.7: Exposing INSPIRE data as Linked Data (and related work)
Linked Data community• Archer, P., Loutas, N. & Goedertier, S.
Cookbook for translating Data Models to RDF Schemas
• Colas, C., Goedertier, S., Kourtidis, S, Loutas, N. & Rubino, F. - Core Location Pilot - Interconnecting Belgian Address Data
• Hyland, B., Atemezing, G. & Villazón-Terrazas, B. - Best Practices for Publishing Linked Data
• ISA Programme – Action 1.1 - Deliverable 3.1 - Process and Methodology for Developing Core Vocabularies
• …
Literature study results
11
Click to edit Master title style
Literature
12
If there is anything relevant missing from our list, please contact us: [email protected]
The list will be provided together with this current presentation.
Click to edit Master title style
Literature study results
Example of project on transformation: Geonovum & TU Delft(Linda van den Brink, Paul Janssen & Wilko Quak)
1. Transforming GML into RDFS/OWL (automated using XSLT)2. Transforming UML into RDFS/OWL (annotating the UML model)
Motivation for the work• Currently service based dissemination of GML structured data
• Semantics for predefined domains and clear use cases, controllable • Not flexible in view of new/revised concepts and relations
• Transformation needed from local sources to INSPIRE data specs• The use of RDF and GeoSPARQL might be complementary• RDF transformation can/should be standardized
13
Click to edit Master title style
Literature study results
14
UML data model
(semantics)
GMLApplication
Schema
RDFS/OWL(ontologies) RDF
ISO 19150
Issues/challenges for the mapping as defined in ISO• Closed world of UML versus
open world of OWL• Connection of concepts in
UML and related concepts in vocabularies
• Modeling conventions and restrictions in UML
Questions tackled in the research and experiments• Is it possible to describe a generic transformation from GML to RDF without knowledge
about the underlying model?• How should geometry be encoded in RDF?• How to transform (automatically) the UML model into RDFS/OWL and can it be
integrated with other ontologies on the web
Click to edit Master title style
Literature study results
The experiment - IMRO• Transformation of GML data to RDFS/OWL (automatically) using GML2RDF
• Elements such as names and descriptions are mapped• Objects, also nested features, data types and properties are mapped• Point and surface features transformed to a WKT serialization
• Transformation of UML to RDFS/OWL (semi-automatically) using ShapeChange• Generation of an OWL vocabulary from GML application schema and UML model• Improved mapping from UML to OWL by adding information (annotation)
• Attributes• Classes
• Open issues• Modifications ShapeChange
• Open-world oriented ontology
• Tackle UML anomalies
15
Appl
icati
on sc
hem
a
met
adat
a
serv
ices
code
list
s
Regi
strie
s
Oth
er
Geonovum Experiments
Click to edit Master title style
Agenda
Introduction
Context
3
2
16
Literature study results
4
5
Sharing your experience – Issues/Challenges?
1
Next steps - Experiments
Sharing your experience – Issues/Challenges?
Click to edit Master title style
Sharing your experience Issues/Challenges?
• From the literature review (so far) and from the discussions in the London workshop several issues can be raised ...1. It seems that there are too many semantics in UML stereotypes that are
not amenable to a generic UML to OWL toolset (Taylor)• Potentially loss of much of the intrinsic OWL capabilities• Hampering the interoperability with other RDF datasets For the transformation of domain models, start over again from scratch for
developing an OWL model?
2. Rules for generation of OWL and the work in ISO/TC 211• Rule-based conversion without harmonization with existing ontologies• Relating it to other OWL ontologies can be done by using RDFS/OWL
mechanisms• Other set of rules exist (e.g. in ShapeChange) Do we need to go for a flat rule-based conversion or not?
18
12/12
Click to edit Master title style
Sharing your experience Issues/Challenges?
• From the literature review (so far) and from the discussions in the London workshop several issues can be raised ...3. Standards for geometry in RDF
• Which standards to use? Debate during the London workshop: GeoSPARQL• Which alternative? Revise GeoSPARQL ?
• In the Dutch experiments GeoSPARQL was used with the WKT serialization• Only Simple Feature Geometry and because WKT is compact• Many vocabularies and extensions exist
4. Other issues ?
19
12/12
LinkedGeoDataBasic GeoGeoRSS
GeoOWLGeoSPARQL
Core LocationNeoGeo
...
What are the key public service and policy areas where INSPIRE-related RDF can be (re)-used
in e-government?
Click to edit Master title style
Agenda
Introduction
Context
3
2
20
Literature study results
4
5
Sharing your experience – Issues/Challenges?
1
Next steps - Experiments
Next steps - Experiments
Click to edit Master title style
Next steps - Experiments
In order to have a common methodology for the experiments some questions have to be answeredA first series of questions / ideas based on the elements of Generic Conceptual Model
1. How do you transform the general Feature model? What are the rules used for transforming spatial object types, attributes (spatial, temporal, locational, metadata, thematic), associations, constraints?
2. Which related concepts from existing ontologies can be re-used?3. Are there any domain independent core vocabularies such as NASA’s sweet
ontology, UCUM that could be used for units, dimensions…4. Which data types will be used for common data types (dates, integers, strings,
etc…)5. How can target classes be identified in UML?6. How closed or open should the final OWL model be?
22
12/12
Click to edit Master title style
Next steps - Experiments
In order to have a common methodology for the experiments some questions have to be answeredA first series of questions / ideas based on the elements of Generic Conceptual Model
7. How to include links to application schema’s, INSPIRE feature concept dictionary, themes, code lists in the OWL schema?
8. How to encode geometry? How to encode raster data?9. Should a distinction be made between enumerations and code lists? Is there a
need to distinguish between different types of code lists (extensible vs non-extensible)
10. How to deal with versioning and lifecycle information in OWL?11. How to deal with the concept of voidability?12. How to identify CRS? (now it is part of the geometry literal in GeoSPARQL)13. How to identify temporal reference systems?14. …
23
12/12
Click to edit Master title style
Literature
24
If there is anything relevant missing from our first list of issues/questions to be tackled please say so
What are the priorities?
contact us: [email protected]
Click to edit Master title style
Next steps
25
WEBINAR II: Guidelines & methodologies 5 & 6 May 2014
1. Highlights
2. Refinement
Follow us on Joinup. We will provide the highlights of this webinar
AND….
At any time, your feedback is very appreciated.
Click to edit Master title style
Next steps
26
Join the collaborative platforms in other areas of Open Source & Semantic Interoperability on https://joinup.ec.europa.eu/
Connecting Geospatial Data
Thank you for your participation!
