The Linked Data Showcase- The ‘Why’ and the ‘What’
Peter Bruhn AndersenLinked Data Architect
Agency for Digitisation (DK)
Ønsker du korrekt bullet,
Why a Linked Data Showcase?
2
Linked Data has clear benefits
• Growing recognition of the value of using: • RDF for semantically enhanced data
• Linked Data APIs as interoperability platform for these data
• A Linked Data environment can give us:
• shorter time from idea to implementation
• quicker adaptability to requirements
• more flexibility to specific user needs
• new ways to combine data easily and on the fly
• i.e. virtual knowledge graphs
• enhanced potential for automation
Common
API
Common
API
Common
API
Common
API
System X System Y System Z
Linked Data has clear benefits
…so why are we not seeing a large and widespread Linked Data environment ?
Where are the useful Linked Data sources ?
Why is it so hard to find solution providers with LD-skills ?
Where are the peer community that can advice others ?W3CRDB2RDFW3CSPARQL
?
The LD-environment dilemma
• Potential users needs solution providers with LD-knowledge
• Potential solution providers needs to see a critical mass of LD-solution customers
The LD-environment dilemma positive loop
Rise in number of
Linked Data sources
Easier for users to find
LD solution providers
More value
to users
Linked Data solutions
have more value to
provides
More users wants
Linked Data solutions
Easier to become part of
the LD-environment
Ønsker du korrekt bullet,
…and what is in it?
7
The Linked Data showcase has two major aims
• One is to plant the seeds that can grow into a Linked Data environment• by Linked Data enabling existing data sources across member states
• by using data sources that are useful for a wider community
• The other aim is to answer questions that arise when entering the Linked Data world.• Potential questions about…
back-end
interoperability layer
privacy or security
user interfaces
Potential questions about……the back-end
• Should we move data to a triple store or can we stay on relational database?• when and how to use RDB2RDF
• describe a best practice for getting data from an RDB to a RDF triple store.
• What are the costs, the risk and the benefits?
Convert betwen
RDB and RDF
- on query
RDB
RDB2RDF
Pure RDF-system
w.Triple store
RDB + Triple store
as copy of RDB
RDB
Potential questions about……the interoperability layer
• How do we choose between the different types of APIs?
• Should we provide more than one type of API? Should one type be mandatory?• i.e. LDP + GraphQL
• Should we redistribute the load between clients and servers?• … by using Linked Data Fragments or…?
Common API
SPARQL
Linked Data Platform
GraphQL + JSON-LD
??
Potential questions about……privacy and security
• Security questions:• How can security be maintained at a high level in a Linked Data environment without
decrementing the overall performance to an unacceptable level?• Which security models and authentication protocols should be implemented and how is this
done?
• Privacy questions:• Are there special requirements for Linked Data to be GDPR compliant? • In what way can Linked Data help to complying with GDPR?
Potential questions about……the user interface
• How do we present dynamically Linked Data?• in an easily understandable, coherent way
• Do we need to change our applications to use LD?
Potential questions about……the user interface
• Should we present new types of data in a spreadsheet like way?
Potential questions about……the user interface
• ..or do we let data reference the preferred way of presentation and build applications dynamically?• i.e. SHACL Shapes pointing to Web Components
The ISA2 Linked Data Showcase is…
• A showcase that intends to initiate a cross-border Linked Data environment• based on base registries • both open and access restricted
• A showcase that will function as a guidance on building Linked Data architecture• for both EU member state organizations and solution providers
• A showcase that takes a step-by-step approach to the challenges• first step: Pilot with Open Data only, minimal development• later: gradually extended in scope and variety • continually supported and updated with new and current technology
Where the LDS is currently
• We are working on:• Reference Architecture
• Focus on lifting existing datasets to RDF-enabled data
• LDS-pilot• Serves as POC for the reference architecture
• LDS Wiki• Intended to function as a guidance on building Linked Data architecture
LDS Reference Architecture Building Blocks
Non-functional aspects
Semantic Standards
Technical syntactic
standards
Service reliability
Availability, and
integrity
Documentation
Data quality
management
Back end
Modelling, collection
and validation
Metadata management
Performance and
scalability
Platform interaction
and APIs
RDB-to-Graph
mechanisms
Backend system
specifications
Front end
Open API
specifications
Serialisations for RDF
Annotation of web
pages
Explorative views and
visualisation of LD
Dashboards
Customisable reporting
Services
M2M publication
services
Aggregation services
Technical formatting
services
M2H publication
services
Content management
services
Platform integration
services
LDS pilot – a proof of concept
• Use case: Implement an uncomplicated but realistic use case based on datasets about legal entities and datasets about spatial data
Business Registry Cadastre Registry
address
location dataname
economic activities
legal form
Current data sources in the LDS-pilot
Spatial Data
Streets and addresses
(Belgium)
Placenames
(Norway)
Addresses
(Norway)
Cadastre
(Czech Republic)
Legal Entities
Legal Entities
(Norway)
Economic entities
(Belgium)
Annual accounts
(Norway)
Economic entities
(Czech Republic)
Data sources that are now lifted to RDF
GoodRelations
Schema.org
Registered Organization
Vocabulary
The Organization
Ontology
Location Core
Vocabulary
Simple Knowledge
Organization System
Dublin Core Terms
Asset Description
Metadata Schema
[ Proprietary Vocabulary ]
Legal Entities
(Norway)
Economic entities
(Czech Republic)
Economic entities
(Belgium)
The LDS Wiki
• Intended to function as a guidance on building Linked Data architecture
• Primarily an internal document to gather the acquired wisdom
LDS - onward
• Core challenges the SEMIC LDS Pilot can provide support for
• Booster packages:
• How to kickstart a LD approach in your administration - How to get from data to Linked Data
• Semantical interoperability:
• Vocabularies – how to find them, build them and use them
• Persistent Identifiers
• SHACL – how and where to use it
• Security and privacy concerns
LDS - onward
• Core challenges the SEMIC LDS Pilot can provide support for
• Technical interoperability:
• Formats for RDF – which one to choose• Linked Data APIs – which type to choose for what purpose
• Reference implementations for LD services
• Facilitating developers
• Security & privacy concerns
• How to handle versioning
• Linked Data Applications – how to build them
LDS - onward
• The Linked Data Showcase has no fixed end• …but curent funding ends a year from now
• The showcase will be better with more participants• …so join us
• …with your data
• …with your knowledge
• …with your questions
Ønsker du korrekt bullet,
Thank you
25