Skip to main content

Sharing the CROWN - Establishing a Workflow from Collection Data to Linked Research Data

Hosting organisations
KHM-Museumsverband and Department of Digital Humanities - University of Graz
Responsible persons
Martina Griesser and Christopher Pollin
Start
End

Context

The Imperial Crown of the Holy Roman Empire is one of the most important symbols of European history. Today it is part of the collections of the Kunsthistorisches Museum in Vienna (KHM). In the course of the so-called CROWN project initiated by the KHM, a comprehensive analysis of the Imperial Crown is being carried out. For this purpose, all components like gems, pearls, plates, etc. will be analysed from a scientific, historical and art-historical point of view. This project, which is currently running until 2024, is thus a truly interdisciplinary endeavour. Research data resulting from the application of highly sophisticated analytical techniques to study manufacturing techniques and materials used, is recorded using The Museum System (TMS) . TMS is a widely used software solution designed for museums. It provides a relational database that can be used for inventorying, documenting, and managing collections.

The CROWN project is faced with an in-depth and complex analysis of a single object. It must fulfil the scientific mandate of a world leading museum and has to be based on the already established TMS for data management. Therefore, it is necessary to implement newly developed workflows in the CROWN project using TMS. Concerning data processing this is a flagship project within the KHM-Museumsverband. It is no longer “usual” collection data that needs to be recorded and managed, but highly specific research data that requires special modelling and representation. As specific expertise is not available at the KHM, the proposed project is performed in cooperation with the Department of Digital Humanities (formerly Centre for Information Modelling, ZIM), University of Graz.

To be usable also for future research, the proposed project is not to develop a ready-made software, but to describe a workflow of a best practice solution to create highly complex, formalised and linked research data in the context of museums according to the FAIR criteria. For this reason, the tools used, TMS or GAMS ( Geisteswissenschaftliches Asset Management System ), are exchangeable, and the workflows developed are transferable to other existing systems.

Main Goal

The objective of the proposed project is to develop and implement a best practice workflow for modelling, transforming, and publishing data from tools like TMS into FAIR research data. Such a workflow does not include “usual” collection data, but goes far beyond that. This is due to the highly specific research domains in the museum context for which there are insufficient templates for acquisition and standardisation.

In the case of CROWN, the relevant research question, i.e. origin and dating of the crown, can only be answered on the basis of a large volume of interdisciplinary investigations and findings. These in turn, i.e. the connection of scientific measurements with historical findings, go beyond the usual structure of data describing objects in collections. In order to formally represent both the model and entities in the data set, a Semantic Web approach is chosen. The RDF data thus generated is based on a domain-specific ontology derived from CIDOC-CRM and Basic Formal Ontology (BFO) , linking its entities to controlled vocabularies and Wikidata . Finally, a web-based proof-of-concept prototype of a user interface adapted to the requirements of the different disciplines involved is developed for the domain described, enabling aggregation, visualisation, exploration and analysis of the processed data.

keywords: museums, semantic web, linked open data, ontology, interdisciplinary research, research data, workflows

Outcomes

The project focused on the development of a workflow for the semantic enrichment and normalisation of research data from the museum system (TMS) using the CROWN project as an example. The workflow includes data cleansing, domain modelling, the generation of RDF data from a TMS export and the creation of a prototype in GAMS.

A particular challenge of the research project was that 2470 objects (as of March 2024; further objects to follow), 669 specific and a further 40 predefined data fields had to be transferred from TMS to RDF. These data fields contain highly specialised information such as the description of drill holes on gemstones (the imperial crown) as well as the shape, surface, colour of gemstones and analytical procedures carried out, such as Raman spectroscopy. As this is an ongoing project, with different departments working together and new data fields being required over time, an expandable workflow is required. The most important areas are described below.

  • Data input: In the TMS system by interdisciplinary teams of scientists, data cleansing and data normalisation.
  • Data transformation: Transformation of TMS export data into RDF.
  • Data modelling: Development of a data model or ontology.
  • Data visualisation: Creation of a prototype in GAMS.

Data input

The previous data collection was successfully cleansed and normalised. In order to ensure further normalised data entry, a handout and training courses on data entry were carried out during the funding period;

Data transformation

The best practices and workflows are documented and published in a GitHub repository . Several Python scripts that convert the CSV export from TMS into RDF form the centrepiece. The mapping is realised by a detailed listing and definition of the UserFields in a Google Spreadsheet, which allows collaborative adaptations of the data fields. Modifications can be made to create similar workflows for other datasets, such as painting data from TMS, to generate semantic research data.

Data modelling

The project-specific ontology crown-ontology.ttl is intended to describe data from museum collections in a structured way. Interoperability is improved through integration with CIDOC CRM. The ontology supports detailed properties of components of the Imperial Crown as well as additional materials, their origin and condition and is open for updates and extensions. The defined workflow makes it possible to use common documentation for both the mapping and the generation of the ontology.

Implemented Python scripts:

  • datafields-to-ontology.py: Uses local JSON files (top-level class + CIDOC mapping) and the data field information defined in Google Spreadsheet to generate the ontology.

Data visualisation

The data was successfully integrated into the staging environment of GAMS and a web prototype for accessing the data was implemented;

In addition, a low-threshold access to the data and its query using GraphDB was demonstrated as part of the best practice to show that RDF data can also be used in other environments.

Links

more CLARIAH-AT projects more DHA projects