Skip to end of metadata
Go to start of metadata

Having converted data on libraries and related institutions to Linked Data in a previous step, we thought it might be a good idea to visualize that data on a map. For now, the map contains German institutions which are only a subset of the institutions described on Instead of slapping all institutions onto the map at once, we decided to use two filters: the library network an institution belongs to and the type of the institution. A SKOS representation of those types is available here . This SKOS is a conversion of the classification used within the German "Sigelverzeichnis" (a central library adress data base). See the ZETA standards documentation for the documentation of this classification and the used notations (under "2. Bibliothekstyp").

How is it done?

The map uses the Google Maps JavaScript API. For performance reasons, the query performed against the lobid SPARQL-endpoint fetches information for all organisations at once (using AJAX):

PREFIX rdf: <>
PREFIX foaf: <>
PREFIX geo: <>
PREFIX org: <>
SELECT ?verbund ?s ?class ?name ?lat ?lon WHERE {
    GRAPH <>
        ?s rdf:type foaf:Organization .
        ?s org:linkedTo ?verbund .
        ?s org:classification ?class .
        ?s foaf:name ?name .
        ?s geo:location ?loc .
        ?loc geo:lat ?lat .
        ?loc geo:long ?lon .

The data is cached on the client side and the actual filtering is done locally by JavaScript. This trades an increase in the time needed for the initial loading for a much better interactivity. The polar opposite would be to generate and execute a SPARQL query every time the filter is changed.

Finding flaws with visualisation

This map shows that one value of data visualization lies in its capacity to point us to flaws and omissions in data sets. Playing around with the map, for many will probably lead to updating their entry in the German Sigelverzeichnis where the underlying data stems from.

Some of the things we have noticed are:

  • German National Library (Deutsche Nationalbibliothek) obviously isn't classified as "National Library" in the Sigelverzeichnis.
  • Only one of the three German central subject libraries is classified as such. Instead you'll find one library in this group which doesn't belong there.
  • Some classification terms aren't used at all:
    • Blinden- und Blindenhörbibliothek / Library for the Blind
    • Verlag / Publisher
    • Verband, Organisation / Association
    • Regionaler Zentralkatalog / Leihverkehrszentrale / Regional Catalogue
    • Fachzentralkatalog / Central Subject Catalogue
    • Serviceeinrichtung / Service Institution
  • Some classification terms are hardly used:
    • Museum
    • Musikbibliothek / Music Library
    • Kinderbibliothek, Jugendbibliothek / Children's Library, Youth Library
    • Ausbildungsstätte / Educational Institution
    • Gefängnisbibliothek / Prison Library
    • Patientenbibliothek / Patients' Library
    • Fahrbibliothek / Mobile Library
    • Ministerium / Ministry
    • Sonstige Einrichtung / Other Organization (Instead many heterogenous institutions are classified as Spezialbibliothek / Speacial Library.)

In short, less than a half of the classification terms are in signicant use as 15 of 27 terms aren't used at all or only to an inconsiderable extent.

Obviously visualizing the data on the map makes evident many potential ways to improve the underlying data. Hopefully it leads some libraries to update their entries in the Sigelverzeichnis or at best to provide this information on their websites using RDFa-enriched HTML so that it is possible to aggregate the most up-to-date data directly from the institutional websites.

Furthermore in our work of SKOSifying the institution type classification, we noticed that the classification as defined in the ZETA standards documentation differs from the possible classes one can choose in the online submission form of the Sigelverzeichnis. The differences are documented here. Hopefully someone from the Sigelstelle will harmonize this.