Debruyne, C., De Leenheer, P., Spyns, P., van Grootel, G., Christiaens, S. (2011) Publishing open data and services for the Flemish Research Information Space. In Proc. of the International Conference on Conceptual Modelling 2011: ER Workshops (ER 2011), LNCS, Springer
Abstract. The Flemish public administration aims to integrate and publish all research information on a portal. Information is currently stored according to the CERIF standard modeled in (E)ER and aimed at extensibility. Solutions exist to easily publish data from databases in RDF, but ontologies need to be constructed to render those meaningful. In order to publish their data, the public administration and other stakeholders first need to agree on a shared understanding of what exactly is captured and stored in that format. In this paper, we show how the use of the Business Semantics Management method and tool contributed in achieving that aim.
DevEX - reference for building teams, processes, and platforms
Publishing Open FRIS Data Services
1. Publishing open data and
services for the Flemish
Research Information Space
Christophe Debruyne, Pieter De Leenheer, Peter Spyns,
Geert van Grootel, Stijn Christiaens
1
2. Outline
Ø Introduction
Ø Problems
Ø Approach: Method and Tool
Ø Demo
Ø Results
Ø Conclusions and Future Work
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.2
3. Introduction
Ø A good overview of a country or region’s science
and technology base is crucial to stimulate
research and innovation
Ø The Department of Economy, Science and
Innovation (EWI) therefore launched the Flanders
Research Information Space (FRIS) program
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.3
4. Introduction: Goals of FRIS
Ø A virtual research information space covering all
Flemish players in the field of Economy, Science
and Innovation
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.4
5. The current portal
researchportal.be
Researchers
Organizations
Projects
Publications
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.5
6. Introduction: Goals of FRIS
Ø Reduce the current administrative burden of
knowledge institutions
Ø Repeatedly report the same information to
different institutions in different formats
Ø Services could automate this process if data would
be accessible in a uniform way
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.6
7. Problem
Ø Two problems:
1. What does the data mean? à Capturing the
semantics of the domain in an ontology
2. How do we appropriately annotate the data
(from heterogeneous) sources?
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.7
8. Problem
Ø Heterogeneous data sources
Ø Knowledge institutions store information in autonomously
developed information systems
Ø European Commission furthermore asks knowledge
institutions to report according the Common European
Research Interchange Format (CERIF)
Ø Knowledge “scattered” across different communities
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.8
9. Problem
Ø CERIF
Ø Created with Entity-relationship model for an
almost unlimited flexibility on roles and
classifications used with entities
Ø Show limitations when it comes to
communicating the knowledge stored with that
model
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.9
10. Problem
Ø To populate the FRIS portal with delivered CERIF
and other data sources, we need
1. A consensus amongst the involved parties – forming
communities – on a common conceptual model (ontology)
2. An easy, repeatable process for validating and integrating
data
3. A means to use that data in a generic way so services could
be built
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.10
11. Approach: Method and Tool
Ø Business Semantics Management
Ø Community driven ontology engineering
Ø Adopts Semantics of Business Vocabulary and Rules (SBVR)
Ø Fact-oriented (like NIAM, ORM)
Ø Communicating (binary) facts
Ø Decision attribute/relation not at design time, but at
implementation time.
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.11
12. Approach: Method and Tool
De Leenheer, P. (2009) On Community-based Ontology Evolution:
Foundations of Business Semantics Management, PhD Thesis, Vrije Universiteit Brussel
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.12
13. Approach: Method and Tool
Ø Business Semantics Glossary
Ø Supports Semantic Reconciliation
Ø Semantic Communities à A group of stakeholders who want
to achieve interoperability
Ø Speech Communities à Subsets of a semantic community
employing a certain terminology
Ø Vocabularies à Sets of facts expressed in a terminology of a
speech community + constraints capturing the intended
universe of discourse
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.13
14. Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.14
15. Approach: Method and Tool
Ø Exporting OWL
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.15
16. Approach: Method and Tool
Ø Exporting OWL
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.16
17. Approach: Method and Tool
Ø Annotating existing data sources
Ø We used D2R Server1, generates a “skeleton” vocabulary
based on a relational database schema
Ø Semantics of the exported ontology are added to the
skeleton file
Ø Result online à http://starpc20.vub.ac.be:2020/
1http://www4.wiwiss.fu-berlin.de/bizer/d2r-server/
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.17
18. Approach: Method and Tool
Ø Demo of Business Semantics Glossary
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.18
19. Results
Ø SPARQL endpoint is running
Feel free to test and use it!
Ø Services built on top of that data
Ø Results submitted elsewhere
Ø For instance,
http://starpc14.vub.ac.be:8080/mashup/person
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.19
20. Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.20
21. Conclusions
Ø A teachable/Repeatable process for capturing
knowledge in an ontology and annotating data
Ø Heterogeneous data sources uniformly accessible
through one ontology
Ø Development of services automating (reporting)
tasks now possible
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.21
22. Future work
Ø Facilitating the annotation process. E.g., transforming Ω-RIDL
annotations into D2R mappings
Ø Lifecycle management
http://www.acsi-project.eu/
Ø Hybrid Ontology Engineering for Linked Data
http://www.oscb.be/
Publishing open data and services for the
Flemish Research Information Space
2/11/11 Pag.22