impact of standards in european open data catalogues. a multilingual perspective of dcat

7/27/2019 Impact of Standards in European Open Data Catalogues. A Multilingual perspective of DCAT

1/27

ePSIplatform Topic Report No. 2013 / 09 , September 2013

1

ImpactofStandardsinEuropeanOpenDataCatalogues.AMultilingual

perspectiveofDCAT

EuropeanPublicSectorInformationPlatform

TopicReportNo.2013/09

ImpactofStandardsinEuropeanOpen

DataCatalogues.AMultilingual

perspectiveofDCAT

Authors:ElenaMontiel-Ponsoda,BorisVillazn-Terrazas

Published:September2013


2/27


2


perspectiveofDCAT

TableofContents

TableofContents......................................................................................................................2Keywords:..................................................................................................................................3Abstract/ExecutiveSummary:..................................................................................................3

1 Introduction............................................................................................................................42 DCATOverview.......................................................................................................................53 DCATcompliantdatacatalogs................................................................................................74 SomeissuesrelatedwiththecurrentuseoftheDCATvocabulary:alanguageperspective12 5 EnrichingRDFvocabularieswithmultilingualinformation...................................................196 Approachesfortherepresentationofculturallyinfluencedelementsinontologies...........217 Conclusionsandrecommendations......................................................................................23AbouttheAuthors...................................................................................................................25References..............................................................................................................................26Copyrightinformation.............................................................................................................27


3/27


3


perspectiveofDCAT

Keywords:

Datacatalogs,DCAT,PublicSectorInformation,Multilinguality

Abstract/ExecutiveSummary:

Severalcountriesfromaroundtheworldareestablishingdataplatforms.WithintheEuropean

Unionmemberstates are settingup officialdata cataloguesas entrypointsto simplifypublic

access to PSI (Public Sector Information) of the country. In this context it is important to

describethemetadataofthesedataportals, i.e.,datacatalogsand allowthe interoperability

amongthose.Totackletheseissues,theGovernmentLinkedDataWorkingGroupisdeveloping

DCAT(DataCatalogVocabulary),anRDFvocabularyfordescribingthemetadataofdatacatalogs.

Thistopicreportdescribestheadvantagesofhavingastandardfordatacatalogsandanalyses

themultilingualperspectiveofDCAT.


4/27


4


perspectiveofDCAT

1IntroductionIn recentyearsdata hasbecomethenewoil. Indeed,just like oil, itneeds tobe discovered,

extractedfromitssources,andrefinedfromtherawmaterialintoproductswithahighadded

value. Following this trend,many national, regional and local governments, aswell as other

organizationsinsideandoutsidethepublicsector,areoperatingdatacatalogswebportals-

thatprovideaccesstomachine-readablepublicdatapublishedbytheseorganizations.Theneed

for a standard format to represent the metadata contained in these catalogs has been

recognized(Maalietal.,2010),asawaytoimproveinteroperabilityandexchangeofdataandin

ordertoavoidcatalogsendingupbeingdatasilos.

Inthisline,theW3CGovernmentLinkedDataWorkingGroup1isdevelopingDCAT(DataCatalog

Vocabulary), an RDFvocabulary designed to facilitate interoperability betweendata catalogs

publishedontheWeb(Maalietal.,2013).DCATwasfirstdevelopedandpublishedbyDERI2and

has seen widespread adoption at the time of this publication. The original vocabulary was

further developed by the eGov Interest Group3, before being brought onto the

RecommendationTrackbytheGovernmentLinkedData(GLD)WorkingGroup.

By using DCAT to describe datasets in data catalogs, publishers increase discoverability and

enable applications easily to consume metadata from multiple catalogs. It further enables

decentralized publishing of catalogs and facilitates federated dataset search across sites.

AggregatedDCATmetadatacanalsoserveasamanifestfiletofacilitatedigitalpreservation.

WithintheEuropeanContext,therearemorethan70officialdatacatalogues4andthereare23

officiallyrecognized languages.In this sense,ourclaim isthatmultilingualism isan important

featurevocabularieshaveto takeintoaccount.This reportanalyzesDCATfromamultilingual

perspectiveandproposestwooptionsforincludingmultilingualisminagivenvocabulary.

1http://www.w3.org/2011/gld/charter

2http://deri.ie/

3http://www.w3.org/2009/06/eGov/ig-charter.html

4

A list of official data catalogues in EU27 member states can be found at http://datacatalogs.org/group/eu-

official


5/27


5


perspectiveofDCAT

2DCATOverviewDCAT is an RDF vocabularywell-suited for representing data catalogs. It defines three main

classes(seeFigure1):

dcat:Catalog,aclassthatdefinesacuratedcollectionofmetadataaboutdatasets.

This class is described by the following properties: title, description, date of formal

publicationofthecatalog,mostrecentdateofmodificationofthecatalog,languageof

thecatalog,a linkto thelicensedocumentunderwhichthecatalogismadeavailable,

the rights underwhich the catalog can be used/reused, the spatial coverage of the

catalog,andalinktothehomepageofthecatalog(seeFigure1).

dcat:DataSet, is defined as a collection of data, published or curatedby a single

agent,whichisavailableforaccessordownloadinoneormoreformats.Ascanbeseen

in Figure1, there are numerous properties that describe this class: title, description,

dateofformalissuanceandmostrecentdateofmodification,auniqueidentifierofthe

dataset, a keyword or tag describing the dataset, the language of the dataset, the

temporal period that the dataset covers, the spatial coverage of the dataset, the

frequencyatwhichthedataset ispublished,and, finally, thelandingpage, i.e.,aWeb

page that can be navigated to in aWeb browser to gain access to the dataset, its

distributionsand/oradditionalinformation.

dcat:Distribution,aclasswhichconnectsadatasettoitsavailabledistributions.

Thelatteraredefinedbypropertiessuchastitle,description,dateofformalpublication

andmost recentdateofmodification,linksto thelicensedocument underwhich the

distributionismadeavailable,aURLthatgivesaccesstoadistributionofthedataset,a

directlinktoadownloadablefileinagivenformat,mediatypeofthedistribution,the

fileformatofthedistribution,andthesizeofadistributioninbytes.

Moreover,throughthedcat:themeproperty,DCATreliesontheclassskos:Conceptfor

classifying its datasets according to a set of domains or categories, which in its turn are

categorizedororganizedinataxonomyusedtorepresentthemes/categoriesofdatasetsinthe

catalog(dcat:themeTaxonomy).


6/27


6


perspectiveofDCAT

Figure1.DCATmainclasses(Maalietal.,2013)

Therestofthedocumentisorganizedasfollows.Insection3weprovideananalysisofsome

datacatalogswhichareannotatedwiththeDCATvocabulary,andwhichcontaindatainseveral

languages.Theanalysisprovidesaninterestinginsightintotheactualuseofthevocabularyand

highlights some issues related with the use of the vocabulary inmultilingual settings or by

publisherswhosemainlanguageisnotEnglish.Theseissuesaresummarized,exemplifiedand

discussed in section 4. Then, in section 5 we present several modeling options for the

enrichmentofRDFvocabularies,such as theDCAT vocabulary, withmultilingual information,

whatwould come to solve some of the problemsof its current use inmultilingual settings.

Finally, we end the report in section6, by presenting approaches for the representationof

culturallyinfluencedelementsinvocabularies.


7/27


7


perspectiveofDCAT

3DCATcompliantdatacatalogs

In order toassess thecurrentuseof theDCAT vocabulary in European public data catalogs,

firstlyweanalyzedindetailseveralcatalogsthatmakeuseofthisvocabulary.Specifically,the

catalogsusedinouranalysisare:

PublicData.euEuropespublicdata5

ThedatacatalogoftheLocalGovernmentofGijn6,inSpain.

Gencat,thedatacatalogoftheRegionalGovernmentofCatalonia7,inSpain.

ThedatacatalogoftheLocalGovernmentofZaragoza8,inSpain.

Theselectionof these catalogswasmotivatedbyaprevious study carried out byDERI first

developers and publishersof the DCAT vocabulary-which resulted in a list of data catalogs

currently making use of this vocabulary. The list, which has been included below for

convenience(seeTable1),wasmadeavailabletousandcontainedninecatalogsintotal.

5http://publicdata.eu/

6http://datos.gijon.es/

7http://www20.gencat.cat/portal/site/dadesobertes

8http://www.zaragoza.es/ciudad/risp/


8/27


8


perspectiveofDCAT

Table1:DatacatalogsDCATcompliant

Catalog Website Type

(national,

local,

regional)

dcat

availabe

at/as

#datasets theme? ke ywords? publisher ex te nt

(spatial)

distribution

format

distribution

size

Semantic

CKAN

http://semantic.cka

n.net

aggregate SPARQL

endpointhttp://seman

tic.ckan.net/

sparql

27592 nope yes...

representedusing

moat:tagged

WithTag

nope nope yes nope

Publicdata.

eu

http://publicdata.eu aggregate A list of RDF files

http://rdf.ope

ndatasearch.

org/

11655 yes, but

each

aggregated

catalog has

its own

hierarchy

i.e. themes

are not

reconciled

yes yes (though

uses

dc:creator)

yes yes nope

Barcelona http://w20.bcn.cat/

opendata/

City RDF dump

http://w20.bc

n.cat/openda

ta/CatalegR

DF.aspx

501 nope yes yes nope yes nope

Gijn http://datos.gijon.e

s

City an RDF file

per dataset

29 nope yes (though

wrongly

uses

dc:keyword)

yes nope yes yes

Catalonia http://dadesobertes

.gencat.cat

Regional RDF dump

http://dades

obertes.genc

at.cat/recurs

os/datasets/

cataleg.rdf

127 yes (though

uses

dc:subject)

yes yes yes yes yes

Balearic

Islandshttp://www.caib.es/

caibdatafront

Regional RDF dump

http://dades

obertes.caib.

es/recursos/

datasets/cat

alog/rdf

28 yes (though

uses

dc:subject)

nope yes (though

wrongly

uses

dc:editor)

nope yes (though

in a wrong

way)

yes (in a

wrong way

though)

Saragossa http://datos.zaragoz

a.es

City SPARQL

endpoint at

http://www.z

aragoza.es/d

atosabiertos/

sparql

271 yes (though

uses

dc:subject)

yes yes(wrongly

uses

dc:publiser

with no

label or any

other

property.

also uses

the non-

existing

dc:editor )

yes yes nope

Fingal http://data.fingal.ie Regional by t hird

party (DERI)

73 yes nope yes yes yes nope

lab.linkedd

ata.deri.ie

http://lab.linkeddat

a.deri.ie/govcat

Aggregate SPARQL

endpoint at

http://lab.link

eddata.deri.i

e/govcat/spa

rql

1710 yes yes yes yes yes yes

Inthefollowing,weprovideabriefdescriptionoftheselectedcatalogsfocusingonthepurposeofthecatalog,languageoftheportal,thelanguageofthedatasets,andthesearchoptions.

ThePublicData.euportal(Figure2)aimstoprovideaccesstoopendatasetsfromlocal,regional,

andnationalpublicbodiesacrossEurope.Apartfrombeingasinglepointofaccesstoscattered

datasetsinEurope,itenablesuserstobrowsedatasetsbyregion,subjectmatterandformat.

AlthoughtheportallanguageisEnglish,wecansaythatthePublicData.eucatalogismultilingual

in thesense that itcontains datasets originating from several European countriesandwhich

containinformation invariousnatural languages.Thedatasets come fromUK,France, Spain,


9/27


9


perspectiveofDCAT

Austria, Denmark, Italy, Czech Republic, The Netherlands, Germany, Russia, Sweden, and

Ireland.Infact,anadditionalbrowseoptionoftheportalisbycountryoforigin(seeFigure3).

Figure2.SnapshopofthePublicData.euportal

Figure3.PublicData.eubrowseoptionbydatasetlanguage

Thenext catalog, thecatalogof theLocalGovernmentof Gijn,a city in thenorthofSpain,

containsaround400datasetsinSpanishmanagedbythelocalgovernment.Themainpurpose

ofthepublicationofdatasetsinRDFistocontributetoimprovecitizenparticipation,promote


10/27


10


perspectiveofDCAT

innovationandgivecompaniestheopportunitytocreateaddedvalue,sothatbothpeopleand

marketplayerscanbenefit frompublic information andoffernewproductsandservices.The

portalisavailableonlyinSpanishandonecanfilterandcustomizethesearchresultsbyusing

theoptions:keyword,categoryandformat,asyoucanseeinFigure4.

Figure4.BrowseoptionsintheportaloftheLocalGovernmentofGijn

Gencatisthethirdcatalogwehavestudied(seeFigure5).Mostofthedatasetsinthisportalof

theGovernmentofCatalonia,inSpain,areinCatalan,theofficiallanguageinthisSpanishregion

together with the Spanish language. The portal is available in English, Spanish and Catalan.

According to theinformation in theportal, this catalog comprisesa databaseof the26,000

official facilitiesofCatalonia, the1,400procedures handled in the Governments officesand

some of its multimedia archives. The browse options are by categories (called tags in the

Englishversionoftheportal),format,anddatasources,i.e.,theorganismsthathavecreatedthe

correspondingdatasets.


11/27


11


perspectiveofDCAT

Figure5.SnapshotofthesearchoptionsintheEnglishversionofGencat

ThefourthandlastcatalogthatwehaveconsideredforthisstudyisthedatacatalogoftheLocal

GovernmentofZaragoza,alsoaSpanishregion(Figure6).TheportalisonlyavailableinSpanish,

andtheinformationinthedatasetsitcontainsisalsoinSpanish.Asintheprecedingcatalogs,

thepurposeofthisportalistoprovideaccesstocitizenstopublicdataaswellastofosterthe

reuse of that information. The browser or search options are categories, type of update

proceedingofthedataset(yearly,quarterly,monthly,daily,instantly),formatandkeywords.

Figure6.PortalofthedatacatalogoftheLocalGovernmentofZaragoza


12/27


12


perspectiveofDCAT

4SomeissuesrelatedwiththecurrentuseoftheDCATvocabulary:alanguageperspective

The next step inour analysiswas to access someof the datasets contained in the different

catalogs,availableintheRDFformatandannotatedwiththeDCATvocabulary,andlookintothe

use they made of the DCAT classes and properties. The main conclusions of this study are

discussedbelow.

1. SomedatasetsarenotusingthelastversionoftheDCATvocabulary.Forexample,the

datasetListdesIFSIenIledeFrancecontainedinthePublicData.eucatalogmakesuseof

the properties dct:creator and foaf:name to refer to the publisher of the

dataset,insteadofthedct:publisherpropertyandfoaf:Agentclassdefinedby

thecurrentversionoftheDCATvocabulary.

AsimilarexampleisfoundinthecatalogoftheLocalGovernmentofGijn.Inthecaseof

adatasetofhostels,wefindthefoaf:Organizationclassinsteadoffoaf:Agent

when defining the publisher of the datasets; or the dc:mediaTypeorExtent

insteadofthedcat:mediaTypedefinedinthecurrentversionofthevocabulary.

2. Some datasets make a free use of the DCAT vocabulary, i.e., they are not fully

compliantwithDCAT.Bythiswemeanthattheyusepropertiesofacertainclassinthe

descriptionofanotherclass.Forinstance,in thesamedatasetmentionedabove from

thePublicData.eucatalog, ListdesIFSIenIledeFrance,thepropertyfoaf:homepage

isapropertyoftheclassdcat:Dataset,i.e.,itisdescribingthedataset,whereasit

should be a property of the class dcat:Catalog, as established by the DCAT

vocabulary.

3. Anotherremarkableaspectoftheanalyzeddatasetsisthattheydonotmakeuseofthe

sameamountortypeofmetadata.Thismaybe,tosomeextent,reasonable,sinceeach

publishermightdecidewhichelementsofthevocabularycovertheneedsofhisorher

catalog.Mostcatalogsmakeuseofthedescriptiveinformationrelativetothedataset,

suchas, title, description, date ofissueordate ofmodification,andalso information

related to the distribution of the dataset (see the descriptive properties of the


13/27


13


perspectiveofDCAT

dcat:Distribution classin section2).However,veryfewcontaininformationof

the Catalog itself, of the Record,or of the themeand theme taxonomy used bythe

catalogordatasetinquestion.

4. WhenaccessingthecodeofthedatasetinRDF,werealizedthatALLcatalogsreusedthe

DCAT vocabularyas it is,i.e.,with thelabels for classesand propertiesin English, as

definedby theauthorsof thevocabulary.NoneofthepublisherstranslatedtheDCAT

vocabularyitselfintoitsownlanguage,evenwhen therealdatao information in the

datasetswasinalanguagedifferentfromEnglish.

This isthecommonchoicewhentheontologyorvocabularyisshareableandvalidfor

different cultures. By this wemean that a certain conceptual organization (i.e., the

classesandpropertiesthatmakeupanontologyorvocabularyandthewayinwhich

theyhavebeenorganized)isuniversal,inthesensethatitdoesnotsolelyreflectthe

needs ofa certain cultureorhowa certain cultureapproachesaparticularly area of

knowledge,butitisvalidortranslatabletoothercultures.Infact,thesetofclassesand

propertiesproposedintheDCATvocabularyaregeneralenoughsoastobeacceptedby

anypublisher.Obviously,itmayhappenthatsomepropertiesthatarerelevantforone

publisher are not relevant for another one. For example, the dcat:language

property defining thedcat:Dataset class is highly relevant for the PublicData.eu

catalog,sinceitcontainsdatasetsinseveralnaturallanguages.Inthecaseofthecatalog

of the Local Government of Zaragoza this is not the case, since all datasets contain

informationinthesamelanguage,Spanish,andthisisnotapropertythatdeservesa

specialmention.

However,whatweobserveintheportalsanalyzedisthatthosepublishersincountries

where English isnot thenative language have provided different translations for the

classesor propertiesof the DCATvocabulary that are shown to the user intheweb

page. For example, in the catalog of the Local Government of Zaragoza, one of the

search options is by Materias (Subjects) or Temas (Topics), corresponding to the

dcat:themeproperty,whereastheLocalGovernmentofGijnhastranslatedthisas

Sectores (Sectors) o Sectores temticos (Domain areas). In the case of Gencat, the

catalogoftheGovernmentof Catalonia,theseareCategories (categories),and inthePublicData.eu,theyaredubbedGroups.


14/27


14


perspectiveofDCAT

It could be said thatmateria, tema, sector, sector temtico, categoria or group are

synonymsortermvariantsthatcanbeassociatedtothesameconcept.Onecouldalso

argue that each publisher can translate the names of thesemetadata as he or she

wishes,as long as thetype of information provided by thedifferentmetadata is the

appropriate one. And we agree with this. Nevertheless, when searching different

catalogsinthesamelanguage,letus saySpanish,onewouldexpectto findthesame

information labeled in the same way. We believe this would avoid confusion and

simplifythesearch.

Forthisreason,andalsofortaskssuchastheautomaticgenerationofwebpagesfrom

ontologiesorvocabularies,wewouldbeinfavorofproposingofficialtranslationsof

theDCATvocabulary,whichcouldbeenrichedwithasmanyvariantsaswishedinorder

tocovertheneedsofallpublishers(seesection5formoredetailsonthis).

5. ThelastissuewhichwecameacrossregardingtheuseofDCATbydifferentpublishersin

Europeisthatthecategorizationtheymakeofthedatasetsisalsodifferent.Theauthors

oftheDCATdonotprescribethetopicsorcategoriesschemathatshouldbe followed

whenusingthisvocabulary.Theyonlydetermine that thepropertydcat:themebe

linked to a skos:Concept, which in its turn be included in a

skos:ConceptScheme. Because of this, each publisher has adopted a different

categorizationortaxonomyofcategoriestoclassifydatasets.Wehadthechancetoask

someofthepublishersandtheysaidthattheyhadsimplyadoptedthecategorization

thatthedifferentpublicbodiesalreadymadeintheirrespectivewebpages.

Bywayofexamplewe includebelow theclassification followed in thePublicData.eu

portal(Figure7),andthetaxonomyusedintheGencatportalfromtheGovernmentof

Catalonia(seeTable2).


15/27


15


perspectiveofDCAT

Figure7.ClassificationofdatasetsonthePublicData.euportal


16/27


16


perspectiveofDCAT

Table2.TaxonomyusedintheGencatportalfromtheGovernmentofCatalonia

Theme Includes Sector ID

Administraci pblica Oposicions, Publicacions, Recerca, estudis i anlisis, Altres administracio

Agricultura,

ramaderia i pescaAgricultura, Foment de la producci, Infraestructures rurals, Pesca.

Aqicultura, Recerca, estudis i anlisis

agr-ram-pesca

Associacionisme i

participaciCivisme, Cooperaci al desenvolupament, Entitats, Equipaments,

Participaci ciutadana, Pau i drets humans, Voluntariat

participacio

Comer i consum Consum comerc

EconomiaAssegurances, Defensa competncia, Deutes i sancions, Pagaments,

Tributs

economia

Educaci i formaci

AMPA, Educaci en el lleure, Educaci infantil, Educaci

secundria obligatria, Estudiar a l'estranger, Formaci adults,

Formaci per a docents, Formaci professional, Idiomes, Material

didctic, Mobilitat educativa, Preinscripci i matriculaci, Provesd'accs, Suport a l'alumnat, Altres, Oposicions

educacio

CulturaArts escniques, Arts visuals, Arxius i biblioteques, Cinema i

audiovisuals, Cultura popular, Lletres, Memria histrica, Museus,

Msica, Patrimoni, Recerca, estudis i anlisis, Altres

cultura

Esports, lleure i ociCaa, Esports, Jocs i espectacles, Nutica i busseig, Pesca, Proves

esportives, Vacances i estades

esports-oci

Indstria i energia Estalvi energticindustria-

energia

Habitatge Compra, Lloguer, Protecci oficial, Rehabilitaci, Altres habitatge

JustciaAcadmies, Associacions, Col

legis professionals, Federacions,

Fundacions, Gestors administratius, Oposicions, Recerca, estudis i

anlisis

justicia

Llengua i

comunicaciOccit,Llenguacatalana,Mitjansdecomunicaci comunicacio

Medi AmbientAigua, Ecologia, Estalvi energtic, Flora i fauna, Sostenibilitat,

Altres

medi-ambient

Mobilitat i transport Trnsit, Transports transport

SalutAssistncia sanitria, Foment salut, Higiene, Recerca, estudis i

anlisis

salut

Serveis Socials Cohesi social, Dependncia, Discapacitat, Altres serveis-socials

Societat, ciutadania i

famlies

Adopci i acolliment, Afers exteriors, Afers religiosos, Dones, Gais,

lesbianes i transsexuals, Gent gran, Igualtat, Immigraci, Infants i

famlies, Joves, Recerca, estudis i anlisis, Altres

societat

Tecnologia. Recerca i

InnovaciFoment, Noves tecnologies, Recerca, Societat de la Informaci,

Telecomunicacions, Innovaci

tecnologia

Territori i paisatge.

UrbanismeBoscos, Costes, Paisatge, Ports

territori

TreballBorsa de treball, Cooperatives, Emprenedoria, Formaci, Igualtat

d'oportunitats, Ocupaci, Oposicions, Relacions laborals, Seguretat i

salut laboral

treball

Turisme Establiments turstics, Foment turisme, Altres turisme


17/27


17


perspectiveofDCAT

Theme Includes Sector ID

UniversitatsFormaci en empreses, Mobilitat educativa, Postgraus, doctorats i

msters, Preinscripci i matriculaci, Proves d'Accs a la

Universitat, Recerca, Altres

universitats

IntheGencattaxonomy, thefirst rowidentifiesthemain topics;thesecond, thesub-

topicsrelatedtothemainones,andthethirdrow,thetermusedtoidentifythetopics

in the web page. At a first sight, one realizes that the categorizations of the

PublicData.euportalandtheonesfromGencatdonotmatchincoverageorgranularity.

Inthecaseofcategorizationsofdatasets,weobservesomegeneralconceptsortopics

whicharepresentinanycategorization,whereasothers,maybemoreculturallybound,

areonlypresentincertaincategorizations.InthecaseofPublicData.eu,theypropose14

groups, such as Agriculture, Fisheries and Forestry or Economy and Industry,

whereas the Gencat catalog proposes 22 broad categories. When comparing the

PublicData.eucatalog to theGencatone, we seethat there is some overlap. Inboth

catalogswe find datasets groupedunder thecategories ofAgriculture, Fisheries and

Forestry Agricultura, ramaderia i pesca; Culture and Arts Cultura;

Environment Medi Ambient; or Health Salut. In other cases, the

PublicData.eucatalogmerges relatedcategories, and theGencat catalog keeps them

apart.ItisthecaseofEducationandCommunicationinthePublicData.eucatalogvs.

Universitats(Universities),Tecnologia,recercaiinnovaci(Technology,researchand

innovation), Llengua i comunicaci (Language and communication) and Esports,

lleureioci(Sports,freetimeandhobbies).Inthissense,wecansaythatthetaxonomy

of categories usedby theGencatcatalog is farmore specific than thePublic.Data.eu

catalog.

DuetothefactthattheGencatcataloghasbeencreatedbytheGovernmentofaregion,

in this case, Catalonia, it contains categories that comply with the purpose of the

catalog,suchasAdministracipblica(Publicadministration).Alsohighlyrelatedwith

thedistinctivefeaturesofthisregion,wefindthecategoryofLlenguaicomunicaci

(Language and communication), which contains the sub-categories Occit, Llengua

catalana, Mitjans de comunicaci (Occitane, Catalan language, Media). Such

distinctionsmayrepresentanissueforportalssuchasthePublicData.euone,whichaim


18/27


18


perspectiveofDCAT

at providing a unified access to datasets across Europe, because they will have to

previously analyze the categorizations made by the different portals to aggregate

datasetsundermeaningfulcategories.Sincenocategorizationortaxonomyisprescribed

bytheDCATvocabulary,afteranalyzingseveralcategorizationsofthissort,atentative

categorizationcouldbeproposed,whichcouldbeextendedoradaptedtocoverspecific

culturalneeds.Forthispurpose,severalrepresentationalapproachescouldbefollowed.

Weprovideasummaryofthisinsection6.


19/27


19


perspectiveofDCAT

5Enriching RDF vocabularies with multilingualinformation

Asmentionedinpoint4of theabovesection,withtheaimofenhancingtheuseof theDCAT

vocabularyataninternationallevel,itwouldberecommendabletoprovidetranslationsofthe

labelsthatdescribetheDCATclassesandpropertiestolanguagesotherthanEnglish.Someof

theadvantagesof havingmultilingual versions of this vocabularywouldbe thatpublishers in

countrieswhereEnglishisnottheofficiallanguagecouldmakeuseofthesedescriptionsintheir

ownlanguage,andtheycouldalsodirectlyreusethesetermsorlabelsin theirportalsorfinal

applications. This would also result in all portals making use of the same terms or labels,

contributinginthiswaytointeroperability.

TheideaofenrichingontologiesandRDFvocabularieswithmultilinguallinguisticinformationis

notnewandhasbeentheobjectofresearchandstudyforadecadenow.Tothebestofour

knowledge, some of the first approaches toenrichontologieswith linguisticdescriptions are

LingInfo (Buitelaar et al., 2006), LexOnto (Cimiano et al. 2007), LIR-Linguistic Information

Repository(Petersetal.,2007;Montiel-Ponsodaetal.,2010)orLexInfo(Buitelaaretal.,2009;

Cimianoetal.,2010).Thesemodelsmainlydifferinthetypeoflinguisticdescriptionstheyaimat

accountingfor.Forinstance,whereastheLingInfomodelfocusedontherepresentationofthe

morphologicalandsyntacticstructuresofthoselabelsortermsdescribingontologyclassesand

properties, the LIR model focused on the representation of term variants and translations.

Currently, researchers in this domain have joined forces and are working towards the

standardization ofamodel thatwill intend tocapturea wide rangeof linguisticdescriptions

relative to ontologies or RDF vocabularies. We are referring to the W3C Ontology-Lexica

Community Group9. This standardization initiative has taken the lemon (LExicon Model for

ONtologies)model(McCraeetal.,2011;http://lemon-model.net/)asbasisforitswork,anditis

evolving itintoamodelwhich,incombinationwiththesemanticinformationcapturedin the

ontology, isaimedat improving theperformanceofNLP (Natural Language Processing) tools,

amongst other objectives. See Figure 8 to gain an impression of the kind of linguistic

descriptionsthatcanbeassociatedtoontologyelementsinthe lemonmodel.

9http://www.w3.org/community/ontolex/


20/27


20


perspectiveofDCAT

Figure8.Overviewofthelemonmodel

Asforthe specific caseof the DCATvocabulary, amodel such as lemon would allow for the

inclusionoftermvariantsindifferentlanguagesfortheclassesandpropertiesofthevocabulary.

ComingbacktothepreviouslymentionedexampleoftheseveraltranslationsinSpanishofthe

dcat:themeproperty(materia,tema,sector,sectortemtico,categoriaorgroup),theycould

allbeaccountedforasvariantsorlinguisticrealizationsoftheproperty dcat:theme. Fora

propertysuchasdct:modified,wecouldhavemorereadabletermsorlabels(lastupdate,

changedate,fechademodificacin,fechadeactualizacin,nderungsdatum,Datumderletzten

Aktualisierung,etc.),whichcouldthenbeusedfortheautomaticgenerationofwebpages.


21/27


21


perspectiveofDCAT

6Approachesfortherepresentationofculturallyinfluencedelementsinontologies

Closelyrelatedwiththeapproachesproposedto enrichontologiesandRDFvocabularieswith

multilingual linguistic information is the issue of capturing culturally-bounded classes and

propertiesinontologies.Asmentionedinissuenumber5,section3,theDCATvocabularydoes

notprescribeanycategorizationortaxonomyofcategoriesorthemesintowhichdatasetscan

beclassified.Inthecatalogsanalyzedwefoundoutthatthecategorizationsofdatasetsshowed

some differences, mainlymotivated by the idiosyncrasyof the catalogs themselves,and the

cultureandlanguageinwhichtheyhadbeendeveloped.Inthissense,itwouldbeadvisableto

proposeataxonomyandanalyzewhichapproachisthemostsuitabletomeettheneedsofmost

(ifnotall)publishers.

Taking into account previous work on ontology localization (Montiel-Ponsoda et al. 2010,

Cimianoetal.2010),weenvisiontwopossibilities:

1. Tomapthedifferentcategorizationsbymeansofamappingmodel

2. Tomaintainonecategorizationandtorepresentculturalissuesinanexternallinguistic

modelorasspecificlanguagemodulesorextensionsintheontology

Thefirstapproachallowsforeachpublishermaintainingitsowncategorization,andallofthem

beingmappedorlinkedtoacentralcategorization(seeFigure9fromMontiel-Ponsoda,2011).

However,themappingestablishmentmaybeatoughtask,andsomescalabilityissuesmayalso

appearasmoreandmoredatasetsusetheDCATvocabulary.


22/27


22


perspectiveofDCAT

Figure9.Mappingmodel

Asforthesecondoption,Figure10,onecategorizationwouldbesharedbyallpublishers,andin

case of cultural issues, these could be kept in the linguistic model, or, if needed, specific

culturalmodulescouldbeproposedtoextendtheoriginalcategorization.Themainadvantage

ofthislatterapproachisthatitcontributestointeroperability,butwithoutforgettingculturally

boundissues.InthecaseoftheDCATvocabulary,wewouldbeinfavorofthislatteroption.

Figure10.Vocabularylinkedtoanexternalmodel

Again,thelemonmodeldescribedinsection4(orthemodelthatwillresultfromtheW3COnto-

LexicaCommunityGroup)wouldcometosolvethemodellingissuesinvolvedinthislattermodel.


23/27


23


perspectiveofDCAT

7Conclusionsandrecommendations

The number of data catalogs in Europe is increasing. Lately, there is a trend in public

administrations (regional, local, national and European) to public government data in data

catalogs.DCAT, a vocabulary for representing metadata of data catalogs, is being developed

withintheGovernmentLinkedDataW3CWorkingGroup.ThankstoDCAT,publishersincrease

discoverabilityandenableapplicationstoeasilyconsumemetadatafrommultiplecatalogs.

Inthisreportwehavepresented:

anoverviewofDCAT;

asetofdataportalsthatrelyonDCATfordescribingtheirmetadata;

ananalysisonhowtheseveralportalsactuallyuseDCAT,whichclassesandproperties

theyuse,and,inparticular,howtheyrepresentthethemesoftheircatalog;

the possible options of enriching DCAT with multilingual information, to be able to

representdatacataloguesindifferentlanguages

Ourmainrecommendationistoconsiderthemultilingualismaspectinanyvocabulary,since,on

theonehand,itmaycontributetoitsglobaladoption,and,ontheother,itmayalsoaddto

interoperability.To this respect wehave proposed lemon, amodel for therepresentationof

linguisticinformationrelativetoanontologyorRDFvocabularythatiscurrentlybeingreviewed

forstandardizationpurposes.

Ideally,multilingualismshouldbeconsideredasearlyaspossible,sothatspecificitiesofcertain

languages could be approachedas soon aspossible. This would also allow fora prescriptive

approach,inwhichpublishersaresaidwhichlabels touseineachcase.However,theprocess

rarely follows this order. As vocabularies gain popularity, their adoption increases and

multilingualneedsappeartosupportinteroperability.Infact,widespreadadoptioncomesfirst,

and,then,onerealizesthebenefitsofthemultilingualaspect.Forthesereasons,modelssuchas

lemon allow to maintain the model or vocabulary as it is, and enrich it withmultilingual

informationatanystageoftheprocess.InthespecificcaseoftheDCATvocabulary,andtaken

intoaccount itsgeneraladoption,thenextstepwould involveananalysisofthecatalogsand

portals that implement it to identify the labels used by the various publishers in different

languages. All those labels, or preferably, the ones that better express the meaning of the

vocabularytermsshouldbecapturedinthelinguisticmodelandrecognizedaspreferredlabels


24/27


24


perspectiveofDCAT

ineachlanguage.Thebenefitofthisapproachisthatthemodelwouldtakeadvantageoflabels

(variantsortranslations)thatarepopularandacceptedbypublishers,andwouldnotimpose

theuseofsomelabelsthatmayendupnotbeingmeaningfulforusers.Themodelwouldalso

leavethedooropenfornewlinguisticneedswithoutinterferingwiththeoriginalvocabulary.

Moreover, we believe that it should be made following a conciliatory approach in which

different options are welcomed and integrated, and in which different communities can

participate in proposing termsand translations in their own languages, thus building it in a

cooperative way. All in all, the enrichment of the vocabulary with multilingual linguistic

information would contribute to a wider adoption and increased understanding and

interoperability.


25/27


25


perspectiveofDCAT

AbouttheAuthors

ElenaMontielPonsoda isLecturerattheUniversidadPolitcnicadeMadrid,inMadrid,Spain,

andmemberoftheOntologyEngineeringGroupatthesameuniversity.ShereceivedherM.A.in

ConferenceInterpretingandTranslation(September2000)byUniversidaddeAlicante,herB.A.

inTechnicalInterpreting(February2003)byHochschuleMagdeburg-Stendal,Germany,andher

PhDonAppliedLinguistics (January 2011) byUniversidadPolitcnica deMadrid.Hercurrent

research activities include, among others: Terminology and Translation in the field of

InformationTechnologyandNaturalLanguageProcessing(NLP),inwhichshehasparticipatedin

differentinternationalprojectsconcerningterminology,ontologiesandmultilingualismandits

applicationto theSemanticWeb.She haspublished thebook"Multilingualism inOntologies.

BuildingPatternsandRepresentationModels",andnumerouspapersinjournals,conferences

andworkshopsintheareasofAppliedLinguistics,SemanticWeb,andNLP.

BorisVillazn-TerrazasisLinkedDataResearcherManageratiSOCO.HeholdsaPhDinArtificial

IntelligencefromUniversidadPolitcnicadeMadrid.HehaspreviouslyworkedasPost-Docat

theOntology EngineeringGroup.Beforehewasa researcherand software developer at the

ResearchInstituteofInformaticsattheUniversidadCatlicaBolivianaSanPablo.Hisresearch

interestsarefocusedonLinkedData,SemanticWebandOntologyEngineering,amongothers.

Hehasparticipated inseveralEuropean researchprojects suchasKnowledgeWeb,OntoGrid,

SEEMP,NeOn,SemsorGrid4Env,PlanetData,andParlance,aswellasinnationalprojectssuchas

Reimdoc, Servicios Semnticos, Plata, Gis4Gov, WebN+1, Buscamedia and Ciudad2020.

Moreover, he was leading the Spanish Linked Data initiatives, such as GeoLinkedData,

datos.bne.es,AEMETLinkedData,andElViajero.Finally,hehaspublishedmorethan40papers

in journals, conferences and workshops, and currently he is actively participating in the

RDB2RDF,andGovernmentLinkedDataW3CWorkingGroups.


26/27


26


perspectiveofDCAT

References

1 Maali,F.&Cyganiak,R.&Peristeras,V.(2010). EnablingInteroperabilityofGovernment

DataCatalogues.ElectronicGovernment10thInternationalConference

2 Maali,F.&Erickson,J.&Archer,P.(2013). DataCatalogVocabulary(DCAT),W3CLastCallWorkingDraft.

3 Buitelaar,P.,Declerck,T.,Frank,A.,Kiesel,M.,Sintek,M.,Romanelli,M.,Sonntag,D.,Loos,B.,Micelli,V.,&Porzel,R.(2006).LingInfo:DesignandApplicationsof aModel for the

IntegrationofLinguisticInformationinOntologies .InProceedingsofOntolex2006.

4 Cimiano,P.,Haase,P.,Herold,M.,Mantel,M.,andBuitelaar,P.(2007).LexOnto:AModelforOntologyLexiconsforOntology-basedNLP.InProceedingsoftheOntoLex07Workshop

attheISWC07.

5 Peters, W., Montiel-Ponsoda, E., Aguado de Cea, G., and Gmez-Prez, A. (2007).Localizingontologies inOWL.InFromtexttoknowledge,thelexicon/ontologyinterface,

proceedingsoftheOntolex07workshop.Busan,SouthCorea.

6 Montiel-Ponsoda,E.,AguadodeCea,G.,Gmez-Prez,A.,andPeters,W.(2010).EnrichingOntologieswithMultilingualInformation.JournalofNaturalLanguageEngineering,17(3),

283-309.

7 Buitelaar,P.,Cimiano,P.,Haase,P.,andSintek,M.(2009). Towardslinguisticallygroundedontologies.InProceedingsofthe6thEuropeanSemanticWebConference(ESWC09),111-

125.

8 Cimiano,P.,Montiel-Ponsoda,E.,Buitelaar,P.,Espinoza,M.,andGmez-Prez,A.(2010). Anoteonontologylocalization .JournalofAppliedOntology,5(2),127-137.

9 McCrae, J., Aguado-de-Cea, G., Buitelaar, P., Cimiano, P., Declerck, T.,Gomz-Prez, A.,Gracia, J., Hollink, L., Montiel-Ponsoda, E., Spohr, D., Wunner, T. (2011). Interchanging

lexical resourceson theSemanticWeb.En LanguageResourcesandEvaluation,46,701-719.

10 Montiel-Ponsoda, E. (2011). Multilingualism in Ontologies. Building Patterns andRepresentationModels.LAP-LambertAcademicPublishing.

.


27/27


perspectiveofDCAT

Copyrightinformation

2013 European PSI Platform This document and all material

thereinhasbeencompiledwithgreatcare.However,theauthor,editorand/orpublisherand/or

anypartywithin theEuropeanPSIPlatformor itspredecessor projects theePSIplusNetwork

projectorePSINetconsortiumcannotbeheldliableinanywayfortheconsequencesofusing

the contentof this document and/or any material referenced therein. This report has been

publishedundertheauspicesoftheEuropeanPublicSectorinformationPlatform.

The reportmay be reproduced providing acknowledgement ismade to the European Public

SectorInformation(PSI)Platform.

impact of standards in european open data catalogues. a multilingual perspective of dcat

Documents