1 of 9

The pains of the young data modeller

​

Intro to APIS Tool Gallery, DHP II – 2021-06-10, “Vienna“

​

Matej Ďurčo, Thomas Wallnig

2 of 9

Data Model/ling

  • conceptualize/formalize the objects and relationships of a domain of discourse / application domain
  • reduction – leave out “irrelevant” aspects
  • abstraction – what are the common categories/classes/types
  • crystallisation/convergence point of
    • information at hand or to be processed, or generaged
    • the technology (to be) used, but mainly
    • the goal, purpose, research question, and certainly also discipline or school of thought
  • no one right model - many valid ways to formalize the domain of discourse

3 of 9

Motivation - Why modelling?

“Generally speaking, the importance of data models and data modelling for the digital humanities can hardly be overestimated. Influential scholars consider this to be the core of DH, because modelling is the point where the humanistic understanding of a fraction of reality meets the technical competence of formal modelling, and, in the best of all cases, new questions and fresh research is made possible (Fotis Jannidis; DH 107f).

​

Formalize, make explicit concepts, entities and relations of given domain of discourse

    • To allow communication among peers/stakeholders
    • To allow technical implementation
    • As base for cross-dataset/project integration => interoperability

4 of 9

Incarnations of a data model

Various representations with differing degree of formalisation:

  • Conceptual or semantic model - verbose description
  • Entity Relationship Model
  • UML Class Diagram

5 of 9

Technical implementation

  • Tables (/Relations) in relational database
  • Internal representation in the application�typically classes with properties and methods
  • Frontend /User Interface
  • API exposing data various serialisations / structured formats

6 of 9

Interoperability

  • Motivation:
    • A lot of data produced in various projects
    • Potentially a lot of overlap
    • But uses different technologies/data models => not compatible

​

  • binary function:
    • interoperable(format/model, tool)

= Can I use a certain tool with data of a given type/format/model?

​

  • interoperable(datasetA, datasetB)

= Do they share the same data model?� = Are the data models aligned/compatible?� = Do they refer to the same entities (with the same identifiers?)�

​

7 of 9

Sample datasets / projects

  • VieCPro
    • The Viennese Court – A Prosopographical Portal (ViecPro)
    • ÖAW Innovationsfond 2020–2022
    • https://viecpro.oeaw.ac.at/
    • Implementiert in APIS - Framework zum Management und Publikation von prosopographischen Daten
  • NAMPI
    • Nuns and Monks - Prosopographical Interfaces
    • ÖAW go!digital Next Generation 2019–2021
    • https://nampi.icar-us.eu/
    • eigene technische Implementierung mit
  • Why these datasets?
    • similar domain of discourse
    • very different approaches to modelling of the information

8 of 9

Next up

  • Today still:
    • Theoretical discussion of the modelling approaches in the two sample projects

​

  • Tomorrow - Hands-On:
    • Introducing some basic concepts
    • Exploration of the two datasets and their semantic interoperability overlaps on various levels.

​

​

9 of 9

Thank you! Questions?

Immediate comprehension questions?

(There will be extensive room for discussion afterwards.)