Julia Pabst, Anna Grönig: Report on the conference – “Data for History 2021” – Day 5 (16 June 2021)

Keynote #2

The fifth day of the Data for History conference started with the keynote Social Semantic Annotation: Linking People and Places in the Pelagios Digital Ecosystem by Rebecca Kahn (University of Vienna). This talk was about the experience from the Pelagios project on building a method and a community in and around the use of linked geo-data for working with historical sources. Rebecca Kahn’s talk focused not on the methods of the project, but on the community and network around it.
Pelagios is a Network that connects a method and a community with digital resources. Part of the project consists of providing the Recogito and Peripleo, a search engine for Linked Data.
Pelagios consists of four different founders, an investigative team and many other people involved from different directions (GLAM, students, etc.) The Network was built on the premise of openness regarding data, technology and the community.

The developers encountered two major problems: first, parts of the data to be linked by Pelagios were not freely accessible and second, there was a lack of authority files. One solution was to redirect Pelagios directly to the freely accessible documents. This openness and interoperability, along with the flexibility and pipeline, made Pelagios a success.Since Linked Data works better with more projects people involved,Pelagios has been collaborating with users since the beginning.

Furthermore, the question of trust in Linked Data and the people behind it was discussed.

Building a Common Conceptualization for Research Data in Historical Sciences

In the next session, titled “Building a Common Conceptualization for Research Data in Historical Sciences”, Francesco Beretta (CNRS/Université de Lyon, LARHRA) and Vincent Alamercery (ENS de Lyon, LARHRA) presented OntoME and the Semantic Data for Humanities and Social Sciences Project.
OntoME is designed and built as an ontology management environment by the digital history research team of the LARHRA Lab and the Data for History Consortium. The idea was to build a tool to help modeling ontologies and aligning them to other ontologies.
The high-level ontology in ontoME is based on DOLCE(+) and CIDOC CRM, including its extensions. Based on this, users can add more specific ontologies in the OntoME environment for their model or create their own ontologies.
Vincente Almacerey demonstrated the different functionalities of OntoME and how to use them to build your own ontology and apply it to your data model, for example, through the API in Protégé.
Francesco Beretta highlighted the importance of the community idea for OntoME as well as the community’s crucial role in creating ideas, concepts, and extensions for OntoME. Discussions within the community about different models and different needs to specific data models are crucial for the further development of OntoME.
For this, the OntoME Team offers OntoME Trainings and Semantic Engineering Workshops and invites researchers and other people to join the OntoME project and its developers community. 

In the following discussion, the benefits of OntoME for the Digital Humanities were addressed. OntoME conveniently provides all needed functions for Ontology Modeling in one tool with the advantage of a focus on modelling one’s specific case and specific catering for one’s project. Furthermore, it was addressed how to deal with conflicting concepts and different ways to resolve problems in OntoME. This part of the session especially revolved around the discussion within the community to debate and accept approaches.

Session 10 – CIDOC CRM on the Test Bench

Raphaëlle Krummeich (IRIHS, Université de Rouen Normandie) presented a new research project in her lecture Work in progress attempting to apprehend photography in CIDOC CRM modelling: theoretical contribution to modelling traces, information and meaning constructions. It addresses the research question of how to model autorships of technical drawings with the DOLCE scheme. The method she described uses an event-based modeling process with CIDOC-CRM entities and properties. As examples for the lecture, Krummeich used two drawings of bridges as sources. Each digital document is modeled as a subclass of Visual item, Physical Human-made thing and Propositional object. 

The second presentation of the session Modelling a local historical social network applying an object-oriented and event-based approach compatible with CIDOC-CRM was given by Claus-Michael Schlesinger (University of Stuttgart). The talk described the application of the CIDOC CRM to network data of members of the Stuttgart School. In particular it was discussed, how people and events were modeled. This would allow for analysis of the dataset with network metrics and to employ filters to highlight sub-structures. A major problem with the CIDOC CRM event-based ontology are the social relations between persons and institutions, which are ontologically insufficiently differentiated. What still needs to be done in the next steps is to obtain more empirical data and implement a data-model in CIDOC CRM.

The discussion was about the distinctiveness of the data. Also, domain specific topics of the source, iterative modeling and cross-project linking were reviewed.The applicability of the models to other areas and the linking of knowledge within a community were also discussed, as well as the question of whether CIDOC CRM is useful for network analysis in the art history was also addressed.

Session 11 – Modelling Practices

In the 11th Session “Modelling Practices”, Ingo Frank (Leibniz Institute for East and Southeast European Studies, Regensburg) presented the project “From Legacy Data to Intertwingled Historical Census and Survey Data: A Case Study in Linked Research Data Modeling between Knowledge Organization and Knowledge Representation”.
The idea of the project is to find a way to pragmatically deal with legacy data, especially in the shape of metadata. The common method of storing metadata in CSV and Excel files renders the connection between research data and metadata difficult. With their project, the focus is to link these data seamlessly, i.e. by using RDF data cubes for seamless interconnection between data and metadata.
The Project integrates data and metadata from different Data Projects of the Leibniz Institute for East and Southeast European Studies (OstData, LamBDA, IOS Portal). The main challenges result from incomplete and inaccurate sources as well as the lack of adequate modelled research data.
Here, besides the advantages of interoperability and distribution, the Linked Data paradigm as a solution for source-oriented modeling and the correct integration of historical sources as research data was highlighted. In this regard, RDF Data Cube Vocabulary is used to move from metadata to model the data itself, i.e. to provide data distributions not only as CSV files, SPSS, or Stata files but also to represent the data as linked data. The whole construct is based on granular source-oriented modeling.

The following discussion focused on the still existing problem of structuring and interpreting the collected data, especially when faced with high-dimensional data.
It was also suggested that it may be useful to share the already tested cross-mapping methodology from this project and to make it available for other projects facing similar problems.

Unfortunately, the second presentation by Christoph Schmidt-Petri, Martin Schäler, Michael Schefczyk, Klemens Böhm, and Jens Willkomm (Karlsruhe Institute of Technology) about “The CHQL Query Language for Conceptual History Using Google Books Ngrams” could not take place.

---

About the authors:

Julia Pabst studies history at the Humboldt-Universität zu Berlin in the master’s program with focus on Digital History.

Anna Grönig has a bachelor’s degree in German language and literature and history and is now studying for a master’s degree in history with a focus on Digital History at the Humboldt-Universität zu Berlin.

---

---