Converting tabular data to RDF: Link data

Sample RDF input

@prefix dcterms: <http://purl.org/dc/terms/> .
@prefix skos:    <http://www.w3.org/2004/02/skos/core#> .

<https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016/cz/500011> a skos:Concept ;
  skos:inScheme <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016> ;
  skos:prefLabel "Želechovice nad Dřevnicí"@cs ;
  skos:notation "500011" ;
  skos:broaderTransitive <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016/cz/CZ0724> .

<https://linked.opendata.cz/resource/ec.europa.eu/eurostat/nuts/2013/UKI42> a skos:Concept ;
  skos:notation "UKI42" ;
  skos:inScheme <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/nuts/2013> .

We can enrich the LAU code list with links to related external data. Doing so enables combining the code list with relevant data, such as additional statistics, to perform more sophisticated queries and analyses. In this tutorial, we show linking of two third-party datasets. The first is an unofficial version of NUTS 2010 in RDF, the second are the Czech regions in RDF maintained by the Czech linked open data initiative.

The first dataset is from 2011 and mirrors the 2010 version of NUTS. In addition to the NUTS code list it provides simplified polygons of the geographic geometries of the NUTS regions and contains links to other datasets, such as the European Environment Information and Observation Network, which can be followed to retrieve more data. Our data links the 2013 version of NUTS, which we can connect with the NUTS regions in this external dataset as their previous versions. To do so, we link the older version from the newer one via the dcterms:replaces property, a pattern we already used previously to capture versioning. We can link the dataset simply by appending NUTS codes to the dataset's namespace IRI:

PREFIX :        <http://example.com/>
PREFIX dcterms: <http://purl.org/dc/terms/>
PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>

DELETE {
  ?lau2 :NUTS3_10 ?nuts3Notation .
}
INSERT {
  ?nuts3 dcterms:replaces ?oldNuts3 .
}
WHERE {
  ?lau2 :NUTS3_10 ?nuts3Notation ;
    skos:broaderTransitive+ ?nuts3 .
  ?nuts3 skos:inScheme <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/nuts/2013> .
  BIND (iri(concat("http://nuts.geovocab.org/id/", ?nuts3Notation)) AS ?oldNuts3)
}

The second dataset contains local aliases for the Czech LAU regions. It provides mostly the same data that the LAU code list has, but it is linked from several Czech demographic datasets, such as statistics on births or immigration. As such, it is an example of a dataset that, from our perspective, does not directly bring new data, but its value lies in being linked from other datasets. Once again, we can link the dataset by simply appending the Czech LAU codes to a namespace IRI:

PREFIX skos: <http://www.w3.org/2004/02/skos/core#>

INSERT {
  ?lau2 skos:exactMatch ?region .
}
WHERE {
  ?lau2 skos:broaderTransitive/skos:broaderTransitive/skos:notation ?nuts ;
  FILTER strstarts(?nuts, "CZ")

  ?lau2 skos:notation ?notation .
  BIND (iri(concat("http://linked.opendata.cz/resource/region/", ?notation)) AS ?region)
}

Sample RDF output

@prefix dcterms: <http://purl.org/dc/terms/> .
@prefix skos:    <http://www.w3.org/2004/02/skos/core#> .

<https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016/cz/500011> a skos:Concept ;
  skos:inScheme <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016> ;
  skos:prefLabel "Želechovice nad Dřevnicí"@cs ;
  skos:notation "500011" ;
  skos:broaderTransitive <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/lau/2016/cz/CZ0724> ;
  skos:exactMatch <http://linked.opendata.cz/resource/region/500011> .

<https://linked.opendata.cz/resource/ec.europa.eu/eurostat/nuts/2013/UKI42> a skos:Concept ;
  skos:notation "UKI42" ;
  dcterms:replaces <http://nuts.geovocab.org/id/UKI12> ;
  skos:inScheme <https://linked.opendata.cz/resource/ec.europa.eu/eurostat/nuts/2013> .

The export of the pipeline up to this point is available here.

← Previous step

→ Next step