Tuesday 4 October 2016

LOD can be hugged by human: data.odw.tw.





LODs are nowadays all in love with machines (try 614 examples represented with 3 event types: schema:CreateAction, schema:OrganizeAction, and schema:PublicationEvent.)  Yet it can be hugged by human through CKAN at data.odw.tw
This is the Prat II of the Story of One Leaf, an implementation of the R4R ontology for reusing digital objects for 843312 reused objectsFor instance, see the leaf Pleione formosana Hayata (台灣一葉蘭) in the reusing context.


The Story of One Leaf/ D Version Examples: data:d2148340 (Pleione formosana Hayata 台灣一葉蘭)
data:d2148340 a data:Reused, r4r:RRObject, dcat:Dataset ;
       r4r:hasProvenance data:p20160530-d2148340 ;
       dc:publisher "中央研究院生物多樣性研究中心"^^rdf:PlainLiteral ;
       dc:source "台灣本土植物資料庫
               (http://taiwanflora.sinica.edu.tw/)"^^rdf:PlainLiteral ;
       dc:date "採集日期:1993-04-25"^^rdf:PlainLiteral ;
       dc:coverage "國家:台灣"^^rdf:PlainLiteral,
               "最低海拔:1650"^^rdf:PlainLiteral,
               "行政區:宜蘭縣大同鄉"^^rdf:PlainLiteral .
  • The case data:d2148340 in details is a data:Reused resource, a basic data component in the current Linked Data in the Open Data Web. It uses the R4R ontology to declare its URI as http://data.odw.tw/record/d2148340, and thus it can be shared later with its refined semantic versions using the same URI (ex. its R1 Version). It has provenance information using r4r:hasProvenance to relate the r4r:RRObject with its provenance information.
  • The data:Reused resource is a dcat:Dataset which is a collection of triples, published and curated by ODW, and available for access or download in many versions and formats.
  • Currently, the data:Reused resources are described with Dublin Core 15 Elements (DC15) in close relation to  their primary data. Here we call them DC 15 Version or D Version. Further semantic refinement is done by the refined versions, defined as data:Refined (we call them R Version, with different R1, R2, R3...) extracting information from the values of DC 15 properties. 



    See the leaf  in a semantically refined context:






The Story of One Leaf/ R Version Examples: data:d2148340 (Pleione formosana Hayata 台灣一葉蘭)
data:d2148340 a data:Refined, r4r:Data, dcat:Dataset ;
        r4r:hasProvenance data:p20160706-d2148340 ;
        txn:hasEOLPage  eol:1134120;
        dct:requires evt84:event-d2148340, evt84:phyCre-d2148340 ;
        dcat:landingPage  r1:r1-r2148340;
        dcat:themeTaxonomy data:Biology .
evt84:phyCre-d2148340 a schema:CreateAction ;
        event:factor dct:PhysicalResource ;
        event:product dwc:PreservedSpecimen ;
        dwc:eventDate "1993-04-25" ;
        skos:inScheme dwc:HumanObservation ;
        skos:scopeNote "specimen collection process" .
  • The data:d2148340 is a data:Reused, but also a data:Refined in different contexts. It is a semantically refined version (a r4r:Data in the R Version) of the data:Reused (a r4r:RRObject in D Version). In other words, data:d2148340 and r1:r1-r2148340 share the same URI according to the definition of the r4r:RRObject. which conveys the relation that the r1:r1-r2148340 (a r4r:Data) is a subclass of the data:d2148340 (a r4r:Object).
  • What distinguishes D Version and R Version most ?
    (1) the former with literal values, the latter with URI resource values.
    (2) the latter heavily relies on the event model (further described by using the former that the subject  dct:requires events to denote its people-place-time relations (extracting values of dc:coverage and dc:date from the former). Ex. the event evt84:phyCre-d2148340 triples as shown in the example.
    (3) the latter based on the curation context (data:ConceptTheme, ex. data:Biology) to enrich the resource with domain knowledge like EOL pages, dwc:PreservedSpecimenl and dwc:HumanObservation.




The following slides are a general overview for data.odw.tw.




Triples of the Girl Lost in Thought for Human:

Version D: represented with DC 15 Elements and Provenance (PROV-O)
Version R(r1): represented with refined spatial temporal information and linked to Wikidata, GeoNames, AAT …)



Popular Posts