the europeana datamodel: a semantic layer on top of cultural heritage objects

28-3-2014

1

The Europeana Data ModelA semantic layer on top of Cultural Heritage Objects

20-12-2013 1Roxanne Wyns - IBW

Roxanne Wyns – LIBIS, KU [email protected]

Content

Overview

• Background

• Europeana – From Portal to Platform

• The Europeana Data Model & The Semantic Web

• EDM basic pattern

• Some applications…

• Semantic enrichment @ Europeana

• Some inspiring examples…

20-12-2013Roxanne Wyns - IBW 2

28-3-2014

2

Background

LIBIS

• Library automation service of the KU Leuven

• Support scientific and public organisations in managing their library,

archival and museum collections

• Center of knowledge & expertise

• Expertise in interoperability standards, semantic enrichment and

multilingualism

• European Best Practice research projects (ICT-PSP)

• Member of several standards and enrichment working groups


Before we move to EDM…


28-3-2014

3

Europeana

What is Europeana?

• An internet portal that acts as an interface to millions of books,

paintings, films, museum objects and archival records that have been

digitised throughout Europe

• A platform for knowledge exchange that promotes collaboration

between librarians, curators, archivists and the creative industries

• A platform for access and reuse of cultural content by creative

industries, research and education (Europeana API)


Europeana as a portal


28-3-2014

4


www.europeana.eu

From portal… [1]

From portal to platform…

Aggregation by Europeana

• Gathering digital content from cultural organisations

• Cross-collection and cross-sectoral (archives, libraries, museums)

• Bring this together on the web by using standardised file and metadata

formats (ESE)

• To facilitate resource discovery!

• Quantity vs. Quality


From portal… [2]

28-3-2014

5

A variety of aggregation models

• Project aggregation (ICT-PSP)

– Dark aggregators

– Aggregators + portal

– Domain, thematic, cross-domain

• National or regional aggregation

– Usually with a portal

– Europeana not the only source

– Domain, thematic, cross-domain

• Institutions


From portal… [3]

The European Library


From portal… [4]

28-3-2014

6

Delivering content

• Source > [Intermediate] > Target

• Source = in-house or standard

• Intermediate = standard format or adaptation of standard

• Target = ESE

• Protocols, tools and formats

– XML or CSV

– HTTP, FTP or OAI-PMH upload (provided by aggregator)

– Ingestion and mapping tools (provided by aggregator)

– OAI-PMH (Europeana)

– …


From portal… [5]

From portal… [5]

ESE - Europeana semantic elements

– Represents lowest common denominator for object metadata

– Basically Dublin Core with some extra Europeana elements

– Forces interoperability

– Convert datasets to a “flat” data representation

– Loss of richness of the original data

– Not adequate to accommodate domain specific requirements

– One digital representation per object record

http://pro.europeana.eu/ese-documentation


28-3-2014

7

Moving towards a platform

��


ESEXML

EDMRDF

A change of strategy?!

• Europeana Strategic Plan 2011 – 2015

– Aggregate: Europeana as a trusted source for European Cultural Content

– Facilitate: Support the Cultural Heritage sector through knowledge

transfer

– Distribute: Make Heritage available to everyone, everywhere, at every

moment

– Engage: Find new ways for people to participate in culture

• New Renaissance Report

• Focus on open data and reuse of content

– Linked Open Data, using Semantic Web Technologies

– Data Exchange Agreement (DEA)


… to platform [1]

28-3-2014

8

Transition into 2014 – Why Europeana and not Google?!

• Shift from portal to platform

• Data quality emphasis

• More data providers

• Value creation for contributing partners (aggregators & content

partners)

• Copyright improvement

• Multilingual

• Thematic focus (e.g. Europeana 1914 – 1918)

Europeana Business Plan 2014 (http://pro.europeana.eu/)


… to platform [2]


… to platform [3]

� Positioning Europeana in research, education & creative industries (tourism, social media, UGC…)

� CEF funding

www.europeana1914-1918.eu/

28-3-2014

9

Focusing on data quality

• Introduce star rating system

• Rights statement � increase reuse potential

• Decent previews

• Persistent links

• Metadata records in EDM! �� data interoperability, semantic enrichment, multilingualism…


… to platform [4]

Delivering content

• National aggregators!

• Source > [Intermediate] > Target

• Source = [in-house] or standard

• Intermediate = EDM extension or application

• Target = EDM RDF

• Protocols, tools and formats

– XML, CSV, API?

– HTTP, FTP or OAI-PMH upload (provided by aggregator)

– Ingestion and mapping tools (provided by aggregator, part of CMS)

– OAI-PMH, SWORD? (Europeana)

– …


… to platform [5]

28-3-2014

10

Europeana Inside aggregation


… to platform [5]

http://www.europeana-inside.eu/


… to platform [7]

Europeana API http://www.europeana.eu/portal/api/console.html

28-3-2014

11


… to platform [8]

Europeana LOD pilot

http://europeana.ontotext.com/

The Europeana Data Model (EDM)

& The Semantic Web


28-3-2014

12

EDM & The Semantic Web [1]

Moving towards the Europeana Data Model

• Based on best practices from the different GLAM domains

• Align the data model to the specific community concerns

• Providing different levels of granularity

• Enable the re-use of existing standards

• Providing the possibility to build domain or sector specific application

profiles on EDM



EDM requirements

• Richer metadata - finer granularity

• Distinguish “provided objects” (painting, book, movie, etc.) from their

digital representations

• Distinguish object from its metadata record

• Allow multiple records for the same object, containing potentially

contradictory statements about it

• Support for objects that are composed of other objects

• Support for contextual resources, including concepts from controlled

vocabularies


Introduction to the Europeana Data Model (EDM) (http://pro.europeana.eu/)

28-3-2014

13


A semantic layer on top of Cultural Heritage Objects

• Provides more context to the metadata

• Allows the representation of specific relationships

– Similarities between objects

– Relationships

– Representations

– Derivations

Goal: to make data available as Linked Open Data for re-use by external

sources




The Semantic Web

Linking Open Data cloud diagram: From Wikimedia Commons ; http://lod-cloud.net/

28-3-2014

14



The Semantic web or Web 3.0

• Concept introduced by Sir Tim Berners-Lee (W3C)

• Proposed as the solution for the current problems in sharing and retrieving relevant data on the current Web where:

- Content is not well structured, has inexplicit semantics, is not interoperable (HTML, URLs to link)

- Expressive questions cannot be asked by the user

- Multiple data queries, human interpretation and knowledge is needed to retrieve relevant and “complete” results

�� Moving from documents to data

The Semantic Web is an extension of the current Web

• It includes semantic information (context and meaning!) in web pages

• This meaning allows both people and machines to better interpret the

data

• It creates links so that a person or machine can explore the web of

”related” data via these links

• These links are at the heart of the Semantic web and are needed for

integration and reasoning of data on the Web = Linked Data



28-3-2014

15

Linked Data principles

1. Use URIs as names for things

2. Use HTTP URIs so that people can look up those names

3. When someone looks up a URI, provide useful RDF information

4. Include RDF statements that link to other URIs so that they can

discover related things

Tim Berners-Lee 2007 – http://www.w3.org/DesignIssues/LinkedData.html




Focus shift

• In the web of documents we have HTTP URIs identifying resources

and links between them, but without context:

– What kinds of resources are 'Louvre.html' and 'LaJoconde.jpg'?

– A machine cannot tell, humans can

Towards a Semantic Research Library. Prof. Dr. Stefan Gradmann (KU Leuven)


28-3-2014

16

Towards a Semantic Research Library. Prof. Dr. Stefan Gradmann (KU Leuven)


• So we add syntax for making statements on the resource using RDF

triples and a schema language (RDFS)

• Extending the Web into the ‘Web of Things’



EDM basis

• OAI ORE for organizing an object’s metadata and digital

representation(s)

• Dublin Core for descriptive metadata

• SKOS for conceptual vocabulary representation

• CIDOC-CRM for event and relationships between objects

• RDF for Semantic Web representation


28-3-2014

17


OAI ORE

• Open Archives Initiative Object Reuse & Exchange

• Defines standards for the description and exchange of aggregations of

Web resources

• For combining distributed resources with multiple media types (text,

images, video, data…) � A “bundle” of an object and its digital

representation(s)



Dublin Core + extra EDM elements

• For the descriptive metadata of a cultural heritage object

• edm:ProvidedCHO is the cultural heritage object which is the subject

of the package of data delivered to Europeana

Properties:

dc:contributor, dc:creator, dc:date, dc:format, dc:identifier, dc:language, dc:publisher, dc:relation,

dc:source, dcterms:alternative, dcterms:extent, dcterms:temporal, dcterms:medium, dcterms:created,

dcterms:provenance, dcterms:issued, dcterms:conformsTo, dcterms:hasFormat, dcterms:isFormatOf,

dcterms:hasVersion, dcterms:isVersionOf, dcterms:hasPart, dcterms:isPartOf, dcterms:isReferencedBy,

dcterms:references, dcterms:isReplacedBy, dcterms:replaces dcterms:isRequiredBy, dcterms:requires,

dcterms:tableOfContents, edm:isNextInSequence, edm:isDerivativeOf, edm:currentLocation…


28-3-2014

18

Simple Knowledge Organisation System

• Solution for converting a “classic” thesaurus or vocabulary managed into

a semantically interoperable format

• Based on the RDF specification

• Ideal for creating multilingual networks of terminologies

• Structured according to the ISO 25964 norm

• Components

Concepts Documented

URIs Semantically related (BT, NT, RT)

Labelled Concept schemes





28-3-2014

19

Vocabularies play an important role in the Semantic Web and Linked Data world

• They are the basic building blocks for linking data

• They help with the interpretation and integration of data between

different datasets

• And so may lead to the discovery of new relationships between

information expressed in a different natural language




CIDOC – Conceptual Reference Model (ISO 21127)

• A formal domain ontology for cultural heritage information

• Describes the things that the cultural heritage sector deals with and how these things relate to each other

• Expressed as an “object-oriented” schema

• An object is described according to a series of event that took place in

its lifetime

– When

– Where

– Who

– What


28-3-2014

20


CIDOC-CRM events


Resource Description Framework (RDF)

• Forms the basis of Semantic web technologies

• Universal language to describe the characteristics of resource on the

web

• Using XML for syntax and URIs for naming

• Makes statements about resources in the form of subject-predicate-object triples

• RDF triples provides a labelled connection using URIs to make it

possible to link data with one another

• In this way a machine is able to find the semantic relations between

data



28-3-2014

21

• The different parts of a triple are

– Subject – the thing being described

– Predicate – a trait, aspect, or property of the thing, which expresses a

relationship between the subject and object

– Object – the thing that is the value of the predicate (trait, aspect or

property) of the object thing

• So in the statement “Mona Lisa was created by Da Vinci”

– Subject – Mona Lisa (La Joconde)

– Predicate – Created by

– Object – Da Vinci

• In terms of representation:

– Subject – must be a URI

– Predicate – must be a URI

– Object – may be a URI or a constant value or “literal‟



KU Leuven University

Flanders

is a

Located in


EDM application


28-3-2014

22

EDM basic pattern


EDM basic pattern [1]

• A data provider submits to Europeana a “bundle” of an object and its

digital representation(s)


28-3-2014

23



Musical Instruments Museums Online (http://www.mimo-db.eu/), Rodolphe Bailly

Using DC as the basis for ProvidedCHO

• Advantages

– Wide spread

– Simple

– Stable

– Cross-domain

• Disadvantages

– Not rich enough

– Lack of structure

– No differentiation between the object itself and its digital surrogate (e.g.

creator, photographer � dc:creator)

– Loss of relationships between different classes of data and events (no

relation between who, where, what, when)



28-3-2014

24

“Proxies”

• Describing the provided object as seen from the perspective of a

specific provider

• Used for

– Connecting duplicates of cultural heritage object descriptions coming from

different providers, each with its own metadata

– For adding Europeana enrichments about a resource

– Keeping each provider’s metadata distinct

– And keeping Europeana metadata distinct from the providers’ metadata





aggregation of DMF

aggregation of Louvre


28-3-2014

25




Hierarchical objects

Let’s have a look at some applications…


28-3-2014

26

PartagePlus record provided as EDM


EDM records [1]

Mapping LIDO2EDM

• Respect for the actual specifications of both models in order to ensure

semantic validity of the resulting EDM

• Only a subset of the (core) LIDO elements are mapped

• When value starts with 'http://' or 'https://' it becomes an 'rdf:resource'

in the EDM record, otherwise it is included as a literal

• In addition an EDM property with the preferred label for the concept or

agent in the language of the metadata records as literal is created


EDM records [2]

28-3-2014

27

Mapping LIDO2EDM

• Qualifying information for agents (dc:creator, dc:contributor), dates

(dc:date), places (dcterms:spatial) is lost

• LIDO-based ingestions would benefit from a full implementation of the

EDM model


EDM records [3]


EDM records [4]

28-3-2014

28

Semantic enrichment @ Europeana

Opportunities and pitfalls


Semantic tagging

• Using the AnnoCultor tool (http://semium.org)

– Interprets values

– Searches for corresponding terms in specialised vocabularies

– Adds links to matching terms (dcterms:spatial = Venise � link to place:

http://sws.geonames.org/3164603/

– Pulls in additional information about

this record (βενετία, velence, венеция,

venice, etc.)

Semantic enrichment @ Europeana [1]


28-3-2014

29


Enriched elements

• Place enrichment (edm_place:*)

– Subset of GeoNames (www.geonames.org)

– Limited to European geographic locations

– Limiting on prefixes "A", "P.PPL", "S.CSTL", "S.ANS", "S.MNMT", "S.LIBR", "S.HSTS", "S.OPRA", "S.AMTH", "S.TMPL", "T.ISL“ (http://www.geonames.org/statistics/total.html)

– Enrichment limited to EDM fields “dcterms:spatial” and “dc:coverage”

– Enrichment rules: exact matching?

– Result: 5.8M objects enriched, provides multilingual search on places http://europeana.eu/portal/search.html?query=edm_place%3A*

Issues?

– Appear to be limited

– But only places in Europe are enriched

– And only for the geographical coverage EDM elements



• Concept (topic) enrichment

– Using GEMET thesaurus (http://www.eionet.europa.eu/gemet/)

– 12 concepts removed to avoid linking with homonyms (e.g. Druck)

– Some WWI battles and the two categories “World War I” and “art” from are taken from Dbpedia

– Enrichment limited to EDM fields “dc:subject” and “dc:type”


– Result: 9M objects enriched, http://www.europeana.eu/portal/search.html?query=skos_concept%3A*

Issues?

– Exact matching not limited to the language of the record (Dutch “Tegel” mapped to the Swedish “Tegel”, meaning brick)

– No suitable multilingual concept thesauri for the cultural domain � drawing

– Noise because of metadata quality (dc:type “photo”, “book”, “video”,…)


28-3-2014

30





28-3-2014

31


• Agent (person) enrichment

– Small set of artists (painters) from Dbpedia

– Enrichment limited to EDM fields “dc:creator” and “dc:contributor”


– Result: 136K objects enriched http://www.europeana.eu/portal/search.html?query=edm_agent%3A*

Issues?

– Quality or structure of provided metadata



• Time period enrichment

– Using Semium time periods vocabulary (http://semium.org/time/)

– Partly automatically generated (3rd quarter of 15th century) / manually generated (Roman empire)

– Enrichment limited to EDM fields:dc:date, dc:coverage, dc:temporal, edm:year


– Result: 13.3M objects enriched

http://www.europeana.eu/portal/search.html?query=edm_timespan%3A*

Issues?

– Some words (qualifiers to dates, e.g. “made”, “printed”…) have removed from fields prior to enrichment, but this is only done for English records

– So again a problem of quality or structure of the provided metadata

– Huge issues with BC dates, but also date ranges (e.g. “1701/1800" is mapped to "1701" only)


28-3-2014

32


Pitfalls

• General problems

– Not enough suitable multilingual sources for the DCH domain

– Automatic enrichment vs. manual enrichment

– Quality of the metadata

• Possible solutions

– Indexing vs. display elements

– Full implementation of EDM

– Further extension of EDM

– Gather basic vocabularies and existing multilingual terminologies

– Provide a platform for contributing to translations and mapping vocabularies

– Collect lists for certain metadata fields with limited amount of values, such as format, language, country, date-time rages,…

– Create awareness on Europeana enrichment!



Opportunities

• Multilingual access to over 28 milj. records

• More enriched elements

• Freely available for re-use (DEA)

• Closer to original metadata thanks to EDM

• Data can be contextualized, semantically linked to other data

• Allows for richer semantic query expansion & cross-collection

browsing


28-3-2014

33

Some inspiring examples…


Examples [1]

www.thepund.it


28-3-2014

34

Examples [2]


Examples [3]


www.researchspace.org

28-3-2014

35

Examples [4]


Questions?

Thank you!

Roxanne Wyns – LIBIS, KU Leuven

[email protected]


28-3-2014

36

Questions?

Thank you!

Roxanne Wyns – LIBIS, KU Leuven

[email protected]

www.libis.be

www.libis.be

Sources

• Europeana portal: http://www.europeana.eu/portal/ ; http://www.europeana.eu/portal/api

• Europeana Professional: http://pro.europeana.eu/– Introduction to the Europeana Data Model (EDM)

– Europeana Data Model (EDM) documentation

– Europeana Buisiness Plan 2014

• SPARQL end-point of data.europeana.eu: http://europeana.ontotext.com/

• Towards a Semantic Research Library: Digital Humanities Research, Europeana and the Linked Data Paradigm, Prof. Dr. Stefan Gradmann (KU Leuven)

• Europeana 1914-1918: http://www.europeana1914-1918.eu/

• DM2E: http://dm2e.eu/

• Pundit: www.thepund.it

• The ResearchSpace: www.researchspace.org

• Poisonous India or the Importance of a Semantic and Multilingual Enrichment Strategy, Marlies Olensky, Juliane Stiller, and Evelyn Dröge

• Semantic enrichment at Europeana – memo, November 4, 2013, Antoine Isaac

• Mapping survey LIDO 2013-10-30, Regine Stein


the europeana datamodel: a semantic layer on top of cultural heritage objects

Presentations & Public Speaking