1 / 40

Making Metadata Work Integrating taxonomies with ESCO and applying the ISO25964 SKOS extensions

http:// www.iskouk.org/events/metadata_June_2014.htm. Making Metadata Work Integrating taxonomies with ESCO and applying the ISO25964 SKOS extensions. Presentation for a workshop of the ISKO-UK, IRSG and DCMI joint meeting by johan.de-smedt@tenforce.com agis.papantoniou@tenforce.com.

lily
Download Presentation

Making Metadata Work Integrating taxonomies with ESCO and applying the ISO25964 SKOS extensions

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. http://www.iskouk.org/events/metadata_June_2014.htm Making Metadata WorkIntegrating taxonomies with ESCOand applying the ISO25964 SKOS extensions Presentation for a workshop of the ISKO-UK, IRSG and DCMI joint meeting by johan.de-smedt@tenforce.com agis.papantoniou@tenforce.com ESCO: European Skills, Competences and Occupations ISO 25964 : Standard for mono lingual and multilingual thesauri SKOS : W3C semantic scheme for simple knowledge Organization Systems Integrating taxonomies with ESCO

  2. EC: European Commission DG: directorate general H2020: Horizon 2020, EC programme for Research and Innovation ESCO main reference • ESCO • https://ec.europa.eu/esco/home • European Skills, Competences and Occupations • DG Employment, Social Affairs and Inclusion and DG Education and Culture of the European Commission develop ESCO in collaboration with stakeholders and with the European Centre for the Development of Vocational Training (Cedefop). An H2020 initiative. Integrating taxonomies with ESCO

  3. EU: European Union CV: curriculum Vitae JV: Job Vacancy/Posting ESCO objective • Provide the EU with a set of cross boarder multi-lingual vocabularies to facilitate EU job market transparency by using the ESCO thesauri as a hub for translating and encoding CV and Job Postings. • A set of thesauri/taxonomies (three “pillars”): • Occupations • Skills and competences • Qualifications (certifications) • Semantic relationships between concepts of the three pillars • Occupations and Skills/Competences: essential and optional skills for an Occupation • Occupations and Qualifications: Qualifications required for an Occupation • Qualifications an Skills/Competences: Skills asserted by a Qualification (certificate) Integrating taxonomies with ESCO

  4. ESCO Current Status • ESCO versions • Version 0 is currently published • Version 1 is currently ongoing and will make partial upgrades (v0.1, v0.2 …) until v1 is established (from end 2013 until early 2017). • Version 0.1 is targeted for publication on 2015-02. • Improvements for 3 sectors: • Agriculture, Forestry and Fishery; • Hospitality and tourism; • Veterinary Integrating taxonomies with ESCO

  5. ESCO v0 on the Web • Homehttps://ec.europa.eu/esco/home Integrating taxonomies with ESCO

  6. ESCO v0 Example • Home • Occupations Integrating taxonomies with ESCO

  7. ESCO v0 Example • Home • Occupations • Navigate ISCO-08occupation groupshierarchy till occupationlevel • Only alphabetic sorting at occupationlevel Integrating taxonomies with ESCO

  8. ESCO v0 Example • Home • Occupations • Navigate ISCO-08occupation groups hierarchy till occupation level • Occupation detail Integrating taxonomies with ESCO

  9. Graphical representation of the v0 model Integrating taxonomies with ESCO 9

  10. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  11. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  12. ESCO v1 General Concept Relationship Guideline (1/3) • Use SKOS broader, narrower, and the transitive closures only for hierarchy in a pillar. • Add group-membershiprelation esco:memberOfGroup Integrating taxonomies with ESCO

  13. SKIP Remarks on iso-thes:broaderInstantial and iso-the:broaderGeneric (both not transitive) • As iso-thes:broaderInstantial only can describe the member to leaf-group relationship, a general member – group relationship requires a dedicated memberOfGroupproperty. • The sub-property relation:iso-thes:broaderInstantialoskos:broaderTransitiveSubPropertyOfesco:memberOfGroupobviously does not work in case the ancestor hierarchy is not a hierarchy of nested groups. • Nested groups or classes cannot be captured fully with iso-thes:broaderGeneric because, as it is a sub-property of (the 1 step) skos:broader, the relation cannot be set among (e.g.) the 2-step group 5 and group 513 (on the previous slide).Therefore, similar to the need for a memberOfGroup property complementing skos:broaderInstantial a transitive subGroupOf property would be needed to complement iso-thes:broaderGeneric. • The general property chain would then be:iso-thes:broaderInstantialox:subGroupOf SubPropertyOfesco:memberOfGroup Integrating taxonomies with ESCO

  14. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  15. ILO: International Labour organization ISCO-88: 1988 ILO Occupation codes ISCO-08: 2008 ILO Occupation codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System ESCO v1 General Concept Relationship Guideline (2/3) • Use skos:mappingRelation (e.g. broadMatch and relatedMatch) for relations between (statement subject and object) concepts • both in different concept schemes or pillars • each representing Occupations, Skills or Qualifications or their grouping ancestors. • For mapping into other classification systems, use • Mapping (e.g. skos:exactMatch, skos:broadMatch, … ) to KOS with overlapping domains • ISCO-88, ISCO-08, ROME/FR, … Integrating taxonomies with ESCO

  16. ESCO v1 EnhancementsMapping relationships Accuracy is sometimes problematic - different scope of concepts in different concept schemes Integrating taxonomies with ESCO

  17. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  18. Tagging possibility – NACE hierarchy Integrating taxonomies with ESCO

  19. EC: European Commission dct: Dublin Core terms EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training NACE: Eurostat economic activity sector coding ESCO v1 General Concept Relationship Guideline (3/3) • Use dct:subject for assigning metadata characterizing ESCO concepts • NACE codes, FoET codes, … • ESCO v1 makes an occupation group hierarchy as a sector breakdown. The occupation group scope can be detailed by one or more NACE codes • Use dct:subject to tag an occupation group with the applicable industry sector codes • NACE rev 2; others can be used when applicable • Other applications: • Use tagging to classify skills and qualifications • Allow for tagging of Skills and Qualifications with FoET • Sector specific skills (NACE), skill typing, label typing • EQF of qualification (certification) level Integrating taxonomies with ESCO

  20. ESCO v1 EnhancementsTagging – approach • Recall ESCO v0 to ESCO v1 migration path (see ESCO status - Version v01 planning) • Improvements for 3 sectors: • Agriculture, Forestry and Fishery; • Hospitality and tourism; • Veterinary • ESCO editorial teams • are organized by sectors (as above) • defining occupation groups per sector Integrating taxonomies with ESCO

  21. ESCO v1 Tagging Occupations with one or more dct:subject ranging over NACE codes Integrating taxonomies with ESCO

  22. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  23. CV: Curriculum Vitae / Resume JP: Job Posting / Vacancy Example of gender specific labels • Requirement • CV are gender specific • JP are gender neural • Solution • Coding uses concept • Rendering filters application specific label Integrating taxonomies with ESCO

  24. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  25. Extending and Structuring ESCO v1 • Provide for dynamic sector specific extensions and specializations • Tooling, knowledge, location, … characteristics • Can be modelled as facets • Structural organization is partially facilitated based on ISO-25964 (http://purl.org/iso25964/skos-thes) • Thesaurus Array of sibling concepts, • Concept group, • Compound terms Integrating taxonomies with ESCO

  26. Example of faceting approach • Language groups • Language usage groups Integrating taxonomies with ESCO

  27. Example of faceting approach • The “ConceptGroup” Languages • Provides a “Languages” facet to the “Foreign language” skill • The group can be extended as needed, depending business target • The concept group is a subset of concepts from a language taxonomy (based on ISO 639) • The “ConceptGroup” Language usages • Provides a “Language usages” facet to the “Foreign language” skill • The group can be extended as needed, depending business target • The concept group is a subset of concepts from the Common European Reference Framework for languages (CERF) Integrating taxonomies with ESCO

  28. Integrating taxonomies with ESCO

  29. Labels for compound (or faceted) concepts • The musician Integrating taxonomies with ESCO

  30. Compound Concept versus Compound Term Integrating taxonomies with ESCO

  31. SKIP Compound Concepts (faceted concepts) • Compound concept or Faceted concept • Compound concepts are not provided for by the ISO 25964 OWL. • Hence the ESCO specific conceptsesco:constructedFromesco:hasFacet • Compound terms and compound equivalence • The use ISO 25964 compound equivalence is not the appropriate choice • Compound terms as provided by ISO 25964 are intended for virtual concepts i.e. concepts not actually present in the taxonomy • The ESCO approach does not require for compound concepts to be generated but allows it. Therefore these compound concepts are not (always) virtual. • The better alternative is: • Directly use SKOS labels on the compound or faceted concepts • Derive relationships between component concept labels and compound concept labels from the compound (faceted) concept construct. Integrating taxonomies with ESCO

  32. Thesaurus Array Organization (1/2) • No demo – still building this Concept: Foreign language Facet 1: Language Facet groupspoken languages: can be combined with language usage (en, fr, de, it, es, …) Facet groupsign languages: cannot be combined with language usage (sign language [sgn]) ====================================================== Facet 2: Language usage Facet grouplanguage usage: all language usage concepts Relationships between facet groups: facet group language usageis a facet for all members of facet group facet group spoken languages. Integrating taxonomies with ESCO

  33. <Thesaurus Array> Organization (2/2) • No demo – still building this Integrating taxonomies with ESCO

  34. EC: European Commission ILO: International Labour organization EQF: EC European Qualification Framework FoET: UNESCO Fields of Education an Training ISCO: ILO Occupation codes NACE: Eurostat economic activity sector coding NUTS: Eurostat EU region codes ROME/FR: Occupation codes extended for France KOS: Knowledge Organization System LOD: Linked Open Data SPARQL: (Graph) query language ESCO v1 approach tohandling ESCO v0 drawbacks • Data model • No methodological support for integrating with other relevant KOS • EQF, FoET, NACE, NUTS, … • ISCO 88 • Classification systems used by (national) Public Employment Services (PES): • E.g. ROME/FR, … • Proprietary extension for lay-out and organization • No use of thesaurus array for grouping, sorting and structuring lower levels • Unqualified use of skos:related • Lack of planning for versioning support • Content • One preferred hierarchy (ISCO 08) • ISCO 08 taxonomy is not the ideal, nor the unique for occupation organization system • No handling of gender specific labels • “Waiter” and “Waitress” • No standard framework for detailing the level of Qualification (e.g. EQF) or its subject Area of expertise (e.g. FoET) • Accessibility • Only a portal with download servicesbut no service API, no LOD, no SPARQL end-point Integrating taxonomies with ESCO

  35. SKIP ESCO Efforts for Accessibility • Service API based on REST/JSON • Can be locally applied and integrated in http based applications • LOD API with RDF/XML, N3, Turtle, JSON-LD distribution formats • Central ESCO deployment • SPARQL end-point for all ESCO content and versioning Integrating taxonomies with ESCO

  36. Problems and Conclusions (1/2) • Top level array organization • The ISO 25964 OWL does not hold a top level array organization. This could easily be remedied by allowing skos:ConceptScheme to be in the domain of iso-thes:subordinateArray • Concept groups holding hierarchies • Neither ISO 25964 nor SKOS provide for properties to detail the top members of a group. Though this could be inferred, a “topMember” list is very efficient. • The benefit of skos:memberList (rdf:List) for sorting concepts and arrays • Avoids the need of a sorting key. Such a key may be problematic (e.g. poly-hierarchy) Integrating taxonomies with ESCO

  37. Problems and Conclusions (2/2) • When mapping concepts in different skos:ConceptScheme, the rdfs:subPropertyOf between • skos:mappingRelation (skos:broadMatch, skos:narrowMatch) properties and the respective • skos:semanticRelation (skos:broader, skos:narrower) properties • Typically gets problematic as it makes a dependency between the hierarchies in the different concept schemes • Proposal • Make the mentioned property hierarchy not a required part of SKOS or SKOS-XL, but give them as a possible SKOS extension. Integrating taxonomies with ESCO

  38. EU: European Union ADMS: Asset Description Metadata Schema DCAT: Data Catalog Vocabulary PROV-O: The PROV(enance) Ontology SKIP Versioning – ideas for next steps • Dataset versioning based • DCAT (http://www.w3.org/TR/vocab-dcat/), • ADMS (https://joinup.ec.europa.eu/asset/adms/home) for dataset and resource versioning • Publication Office of the EU • Concept and Label versioning • PROV-O (http://www.w3.org/TR/prov-o/), ADMS • skos-history (https://github.com/jneubert/skos-history) • Versioning ontology for thesaurus or dataset versions • https://github.com/JohanDS/Dataset-versioning--for-KOS-data-sets Integrating taxonomies with ESCO

  39. SKIP Versioning taxonomy publications and taxonomy dependencies • Per thesaurus, one datasetwith fixed URI manages thetaxonomy dataset releases, LOD access i.e. the VersionHistorySet • Each thesaurus datasetrelease is detailed by aDCAT dataset recordi.e the VersionHistoryRecord • A released thesaurus datasetholds an identifier (a URI) for its VersionHistoryRecord and to the VersionHistorySet Integrating taxonomies with ESCO

  40. Thank you • Q&A Integrating taxonomies with ESCO

More Related