Data model for 'Objecten' data

[sdevalk]

@wouteraj As discussed during our meeting, last Tuesday:

We are working on a data model for the 'Objecten' data of SolidCBS. We use CEST's Basisregistratie for this. The Basisregistratie describes the data that collection managers should register for each heritage object in their collection(s).

Unfortunately, there is no Linked Data model of the Basisregistratie. So we have to create our own - and we're not fully finished with this yet. Our model uses Schema.org as vocabulary. Schema.org is not fine-grained, but it supports the general findability of data. The model is intended for publishing data (for use by others), not specifically for storing data in a collection management system, such as SolidCBS.

The model consists of two types of resources:

The first one is a 'Heritage Object'. This resource describes e.g. the context, contents and structure of a heritage object. For example: the title and creator of a painting, the size of a 'bidprentje' or the person depicted on a photograph.

The second one is a 'Digital Object'. This resource describes the digital representation of a heritage object. For example: a picture of a painting or a scan of a 'bidprentje'. Given SolidCBS's initial scope, the digital objects will be images (i.e. not text files, video files or audio files). The resource consists of a handful of properties, notably the URL to the digital object itself and the license for using the digital object.

Applied to the 'Heemkundevereniging Het Lage Land' in SolidCBS, with fictional data, the model could look like this (in Turtle). The example describes a photograph and its digital version. I've inserted various points of discussion. I've also inserted group names, in comments. These correspond to the field groupings of the Basisregistratie.

@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix schema: <http://schema.org/> .

# Heritage object: description of a 'real' heritage object
<https://data.hetlageland.org/heritage-objects/1234>
    ########################################
    # Group: "Identificatie"
    ########################################

    # Field: "Type"
    # To discuss: which types should we (initially) support? For example: schema:Photograph, schema:VisualArtwork, schema:Drawing
    rdf:type schema:Photograph ;

    # Field: "Objectnaam"
    schema:additionalType <https://data.cultureelerfgoed.nl/term/id/cht/1e0adea5-71fa-4197-ad73-90b706d2357c> ;
    # To discuss: the type of the external resource can be anything. Should we make this explicit? E.g.:
    # schema:additionalType [
    #     rdf:type schema:DefinedTerm ;
    #     schema:name "bidprentjes" ;
    #     schema:sameAs <https://data.cultureelerfgoed.nl/term/id/cht/1e0adea5-71fa-4197-ad73-90b706d2357c>
    # ] ;

    # Field: "Objectnummer"
    schema:identifier "SK-A-1115" ;

     # Field: "Titel"
    schema:name "De Slag bij Waterloo"@nl ;

     # Field: "Korte beschrijving"
    schema:description "De slag bij Waterloo, 18 juni 1815"@nl ;

    # Field: "Collectie"
    schema:isPartOf <https://data.hetlageland.org/datasets/1234> ;

    # Field: "Bewaarinstelling".
    # To discusss: should we use this property? It is proposed, not official. Or should the 'schema:maintainer' be part of the dataset description?
    schema:maintainer <https://data.hetlageland.org/> ;

    ########################################
    # Group: "Vervaardiging"
    ########################################

    # Field: "Vervaardiger"
    schema:creator <https://data.rkd.nl/artists/63369> ;
    # To discuss: the type of the external resource may not be a schema:Person or schema:Organization. Should we make this explicit? E.g.:
    # schema:creator [
    #     rdf:type schema:Person ;
    #     schema:name "Jan Willem Pieneman" ;
    #     schema:sameAs <https://data.rkd.nl/artists/63369>
    # ] ;

    # Field: "Plaats"
    schema:locationCreated <http://www.wikidata.org/entity/Q33432813> ;
    # To discuss: the type of the external resource may not be a schema:Place. Should we make this explicit? E.g.:
    # schema:locationCreated [
    #     rdf:type schema:Place ;
    #     schema:name "Delft" ;
    #     schema:sameAs <http://www.wikidata.org/entity/Q33432813>
    # ] ;

    # Field: "Materiaal"
    schema:material <http://vocab.getty.edu/aat/300015050> ;
    # To discuss: the type of the external resource can be anything. Should we make this explicit? E.g.:
    # schema:material [
    #     rdf:type schema:DefinedTerm ;
    #     schema:name "olieverf" ;
    #     schema:sameAs <http://vocab.getty.edu/aat/300015050>
    # ] ;

    # Field: "Einddatum"
    # To discuss: Schema.org doesn't have properties for the 'Begindatum' and 'Einddatum' of creation. How to handle this?
    schema:dateCreated "1824-07-24" ;

    # To discuss: rather than using the properties above we can also use actions (see also 'Group: Verwerving'):
    # schema:potentialAction [
    #     rdf:type schema:CreateAction ;
    #     schema:agent <https://data.rkd.nl/artists/63369> ; # Field: "Vervaardiger"
    #     schema:location <http://www.wikidata.org/entity/Q33432813> ; # Field: "Plaats"
    #     schema:startTime "1824-02-12" ; # Field: "Begindatum"
    #     schema:endTime "1824-07-24" # Field: "Einddatum"
    # ] ;

    ########################################
    # Group: "Voorstelling"
    ########################################

    # Field: "Onderwerp"
    schema:about [
        rdf:type schema:DefinedTerm ;
        schema:name "veldslagen" ;
        schema:sameAs <http://vocab.getty.edu/aat/300185692>
    ] ;

    # Field: "Locatie"
    # To discuss: or use 'schema:contentLocation'?
    schema:about [
        rdf:type schema:Place ;
        schema:name "Waterloo" ;
        schema:sameAs <http://www.wikidata.org/entity/Q31579578>
    ] ;

    # Field: "Persoon of instelling"
    schema:about [
        rdf:type schema:Person ; # Or: schema:Organization
        schema:name "Arthur Wellesley of Wellington" ;
        schema:sameAs <http://www.wikidata.org/entity/Q131691>
    ] ;

    # Field: "Gebeurtenis"
    schema:about [
        rdf:type schema:Event ;
        schema:name "Slag bij Waterloo" ;
        schema:sameAs <http://www.wikidata.org/entity/Q48314>
    ] ;

    ########################################
    # Group: "Verwerving"
    ########################################

    # To discuss: can we register this information, but not publish it?
    # It may contain sensitive data, e.g. prices and names of persons. How to handle this?
    schema:potentialAction [
        # Field: "Type"
        # To discuss: is there a specific class for denoting a 'Verwerving' in Schema.org? E.g. schema:TransferAction, schema:TradeAction?
        rdf:type schema:Action ;

        # Field: "Methode"
        schema:additionalType <http://vocab.getty.edu/aat/300138913> ;
        # To discuss: the type of the external resource can be anything. Should we make this explicit? E.g.:
        # schema:additionalType [
        #     rdf:type schema:DefinedTerm ;
        #     schema:name "schenkingen" ;
        #     schema:sameAs <http://vocab.getty.edu/aat/300138913>
        # ] ;

        # Field: "Datum"
        schema:startTime "1986-06-17" ;

        # Field: "Datum" (same value as 'schema:startTime')
        schema:endTime "1986-06-17" ;

        # Field: "Bron"
        schema:agent [
            rdf:type schema:Person ;
            schema:name "Gijsbert van Loosdrecht"
        ] ;

        # Field: "Aankoopprijs"
        schema:price 1000 ;

        # Field: "Valuta aankoopprijs" (ISO-421)
        schema:priceCurrency "EUR"
    ] ;

    ########################################
    # Group: "Afmetingen"
    ########################################

    # To discuss: Schema.org doesn't support all dimensions of the 'Basisregistratie', such as 'scale' and 'diameter'. How to handle these?
    # Alternatively, we could use a generic 'schema:size' property, with a textual string (e.g. 'hoogte: 56,7cm; breedte: 82,3cm')

    # Field: "Dimensie": "hoogte"
    schema:height [
        rdf:type: schema:QuantitativeValue ;
        schema:unitCode: "CMT" ; # Field: "Eenheid" (UN/CEFACT Common Code)
        schema:value: 567 # Field: "Waarde"
    ] ;

    # Field: "Dimensie": "breedte"
    schema:width [
        rdf:type: schema:QuantitativeValue ;
        schema:unitCode: "CMT" ; # Field: "Eenheid" (UN/CEFACT Common Code)
        schema:value: 823 # Field: "Waarde"
    ] ;

    # Field: "Dimensie": "diepte"
    schema:depth [
        rdf:type: schema:QuantitativeValue ;
        schema:unitCode: "CMT" ; # Field: "Eenheid" (UN/CEFACT Common Code)
        schema:value: 25 # Field: "Waarde"
    ] ;

    # Field: "Dimensie": "gewicht"
    schema:weight [
        rdf:type: schema:QuantitativeValue ;
        schema:unitCode: "KGM" ; # Field: "Eenheid" (UN/CEFACT Common Code)
        schema:value: 1200 # Field: "Waarde"
    ] ;

    ########################################
    # Group: "Overig"
    ########################################

    # To discuss: Schema.org doesn't have properties for denoting the 'Toestand' and 'Huidige standplaats' of things.
    # How to handle these? Perhaps omit these properties for the time being?

    # For linking the heritage object to the digital object
    schema:mainEntityOfPage <https://data.hetlageland.org/digital-objects/1234>
    # To discuss: the type pointed to in 'schema:mainEntityOfPage' can be anything. Is it useful to have a 'convenience property' to the image? E.g.:
    # schema:image <https://data.hetlageland.org/digital-objects/1234> .

# Digital object: description of a digital representation of a heritage object
<https://data.hetlageland.org/digital-objects/1234>
    ########################################
    # Group: "Beeldmateriaal"
    ########################################

    # Field: "Type"
    rdf:type schema:ImageObject ;

    # Field: "URL"
    schema:contentUrl <https://upload.wikimedia.org/wikipedia/commons/thumb/4/48/De_Slag_bij_Waterloo_Rijksmuseum_SK-A-1115.jpeg/1920px-De_Slag_bij_Waterloo_Rijksmuseum_SK-A-1115.jpeg> ;

    # Field: "Licentie"
    schema:license <https://creativecommons.org/publicdomain/zero/1.0/deed.nl> ;

    # For linking the digital object to the heritage object
    schema:mainEntity <https://data.hetlageland.org/heritage-objects/1234> .

What do you think? Can SolidCBS support such a model?

@ddeboer @EnnoMeijers @pietervaneverdingen: feel free to comment!