@openaire_eu
OpenAIRE infrastructureSemantic services in EOSC (EUDAT)
Pedro PríncipeUniversity of Minho
EUDAT ConferenceJan. 23, 2018
2
OpenAIRE’s e-infrastructure
Commons
OpenAIRE Guidelines for
Content Providers
Literature broker service
and the Dashboard for
Content providers
OpenAIRE needs for semantic services
INTEROPERABILITYis the key
OpenAIRE’s e-infrastructure Commons
Publications repositories
Research Data repositories
CRIS systems
Registries(e.g. projects)
OAJournals
SoftwareRepositories
Validation
Cleaning De-duplication
EnrichmentBy inference
Funders, research admins, research communities• Research impact
• Project reporting and monitoring
• Open Access trends
Content providers• Repository validation
• Repository notification broker
• Repository analytics and usage stats
Researchers• Claim publications, datasets, software
• Deposit publications, datasets, software
• Search & browse: interlinked publications, datasets, projects
• Open Access & DMP Helpdesk
• End-User feedback
CONTENT PROVIDERS
INFO SPACE SERVICES
KEY STAKEHOLDERS SERVICES
Project initiative
FunderFunding
Result
Publication Data Software
Organization
GUIDE
LINES
TERMS
OF USE
Guidelines for Data Providers
Literature Repositories
1Data Repositories
2CRIS-CERIF
3
5
https://guidelines.openaire.eu
https://guidelines.openaire.eu
Software Repositories
Catch-all Repositories
CRIS
Data Repositories
Catch-all Repositories
Institutional & thematic
repositories
RESEARCH LITERATURE
Thematic Repositories
Institutional Repositories
E-journals
RESEARCH SOFTWARE
RESEARCH DATA
RESEARCH INFORMATION
FROM Guidelines for Data ProvidersTO Guidelines for Open Science Content Providers
8
https://guidelines.openaire.eu
https://guidelines.openaire.eu
• Funder perspective• Link funding information with research output
• Author and Reader perspective• Link authors and contributors with their research output and ease name disambiguation
• Service provider perspective• Avoid overloading of oai_dc metadata• Make maintenance and mappings of controlled vocabularies easier by help of identifiers• Make identification of resources easier (e.g. for TDM)• Improve alignment with other regional repository networks
• Agree on a shared set of metadata properties and controlled vocabularies
• Allow for region specific extensions
• Examples: LA Referencia, JAIRO (Japanese Institutional Repositories Online)
Upgrade needs from different perspectives
11
Application Profile Overview
• Build an application profile based on established and widely used metadata schemes in repositories• Dublin Core and DataCite v4.1
• Allow for additional properties when needed
• Align with other repository networks
Approach
• Re-use and adaptation of controlled values used in theDataCite schema• E.g. identifier types, role types, relation types
• Controlled Vocabularies defined by the COAR community• E.g. resource types, access rights, version types
• Controlled Vocabularies defined in OpenMinTeD Guidelines• E.g. licenses
Controlled Vocabularies
14
• General aspects• Unique identification of vocabulary concepts• Improved granularity• Multilingual Support• Implemented in SKOS
• Resource Types• http://vocabularies.coar-repositories.org/documentation/resource_types/
• Access Rights• http://vocabularies.coar-repositories.org/documentation/access_rights/
COAR Resource Types and Access Rights
15
Concepts in the Resource Type Vocabulary v1.1
16
OpenAIRE Validator – validator.openaire.eu
17
Test compatibility against OpenAIRE guidelines and
register new repositories
The OpenAIRE enriched information graph offers a great opportunity for repositories to improve their collections…
Literature broker service
THE CHALLENGE
•Enrichment is straightforward• Harvesting from repository and return to repository its records if they
have been “enriched” by deduplication and/or inference
•Addition is less obvious• Based on relationships, in turn identified by inference algorithms
• Must be augmented with notion of “trust” to enable “tuning” options in order to reduce false positive notifications
Literature broker service
19
OpenAIRE Broker sketch
OpenAIRE
Notification Broker
OpenAIRE Information Space
Graph(deduplication,
Inference,
Aggregation)
…
SubscriptionsPotential
Notifications
subscribe
notifyrepository
admin
OpenAIRE Data
Sources
Identifying “events”
relevant to repositories
(enrichments & additions)
Sending
events
Delivered
Notifications
Event (potential notification):
• Message
• Topic
• TargetRepository
• Trust
Repositories can subscribe to the service and receive notifications about records of potential interest and specify
• what metadata fields they would like to be notified of• how to be notified.
The service can notify the repositories in different ways• via custom (OpenAIRE defined) repository APIs for metadata ingestion • via email to the repository managers and via web interface.
Subscription & notification
21
Broker services available via a specific dashboard for content providers…
one stop shop for OpenAIRE data providersfor friends… “the repository managers dashboard”
Dashboard for content providers
Validate
Validation History
Collection Monitor
Enable usage stats
Views and downloads
Events
Enrichments
Notifications
SOURCESRegister
Update
COMPATIBILITY
CONTENT
METRICS
24
25
3. CONTENT >> events & notifications
26
3. CONTENT
27
3. CONTENT >> enrichments
As requested (some ideas):
Opportunities to…
• Improve repositories interoperability.
• Increase metadata quality in repositories.
• Expand the potential of the OpenAIRE information graph.
OpenAIRE needs for semantic services
28
www.openaire.eu
@openaire_eu
facebook.com/groups/openaire