From Data Acquisition to Data Publication: Quality Control, Tool Assessment, and Methodological Challenges in Digital History

Digital tools and AI-assisted methods are increasingly shaping historical research. From OCR and annotation to corpus analysis and the application of generative AI systems in analysis or dissemination, these technologies open up new possibilities for generating and augmenting historical knowledge. Even after a few decades of experience, the combination of hermeneutic and computational approaches is far from straightforward, and researchers are often confronted with new methodological and practical challenges such as (but not limited to):

• Data acquisition: How can processes of datafication and data-collection lead to reliable and accurate data?
• Data modeling: Which data model is best suited to a particular research question and dataset, and how can expressiveness and interoperability be optimally balanced?
• Data analysis: Are the available tools sufficiently powerful for the chosen model, and do they perform as effectively on one’s own data as reported in the literature?
• Evaluation: How can results obtained from a specific tool be critically assessed? What roles do data quality, uncertainty, and methodological transparency play in a given research context?
• Data exchange and reproducibility: How can research results be published and shared in ways that foster reuse and conserve transparency? To what extent are digitally supported methods transferable and reproducible across different datasets, and under which conditions?

The Digital History Autumn School 2026, organized in the framework of the Joint Lab Cultural Data in a Digital World (Herder Institute, Marburg https://www.herder-institut.de/de/start, in close collaboration with the data competence center – Humanities Education in Research, Data, and Methods – HERMES https://hermes-hub.de/), addresses these questions from both methodological and practical perspectives. Together with experts in Digital Humanities and Digital History, participants will explore criteria for the selection, application, and evaluation of digital tools and methods and will have the opportunity to discuss them in the context of their own research projects.

• Selection and critical assessment of digital tools
• Data preparation and data quality practices
• Evaluation methods in the Digital Humanities
• Transparency and reproducibility of digital methods
• Dealing with uncertainty, ambiguity, and incompleteness in historical data
• Source criticism, provenance, and the documentation of data transformation processes

In addition to thematic workshops, the Autumn School will provide opportunities for exchange both with the experts and among the participants. Planned activities include interactive formats such as peer-support sessions, including opportunities for participants to present and discuss their own research data, methods, and research questions as well as networking events.

Contact:

Dr. Cristina Vertan

herbstakademie@herder-institut.de

Program

MONDAY, SEPTEMBER 28
01:00 p.m. Registration / Lunch
02:00 p.m. Opening remarks / Introduction / Short presentation of participants
03:00 p.m. WORKSHOP 1:
How to Use GenAI in (Digital) History and Get Away with It. Evaluating OCR for Historical Documents and the Critical Role of Meaningful Metrics (Host: Ole Meiners)
04:30 p.m. Coffee break
05:00 p.m. Case Studies: Bias and Beyond – Project Presentations
(Host: RTG Marburg)

TUESDAY, SEPTEMBER 29
09:45 a.m. Data Carpentry 1: „Open Refine“ Part I (Host: Julia Tolksdorf and Robert Zwick)
10:45 a.m. Coffee break
11:00 a.m. Data Carpentry 1: „Open Refine“ Part II (Host: Julia Tolksdorf and Robert Zwick)
12:15 p.m. Lunch
01:00 p.m. Workshop Peer-based World Café / review of participants projects / data (Host: Alexandra Büttner, Anna Schlander, Johanna Konstanciak, Ruth Reiche, Andrea Rapp)
03:00 p.m. Coffee break
03:30 p.m. WORKSHOP 2:
Do I need a Part-Of-Speech annotation in my data? The role of linguistic preprocessing (Host: N. N.)

EVENING EVENT AT FORSCHUNGSZENTRUM DEUTSCHER SPRACHATLAS
Venue: DSA, Pilgrimstein 16, 35037 Marburg
06:00 p.m. KEYNOTE by Michael Piotrowski (University of Lausanne)

Michael Piotrowski (University of Lausanne)

Models and Narratives: Computational and Historical Thinking in Dialogue

“Historical Thinking (HT) and Computational Thinking (CT) are both established terms, and especially in a pedagogical context they may appear to be of the same order—particular ways of thinking that students need to acquire—but also very different, if not opposed, due to their association with disciplines that are also commonly perceived as being on opposing ends of the spectrum: whereas HT, at its core, embraces ambiguity, complexity, contingency, and interpretive judgment, CT seeks precision, replicability, and generalizable solutions. There is rarely a single correct answer in historical research: evidence is partial, contradictory, and meaningful only in context. Algorithms, on the other hand, are by definition predictable (in a formal sense). Conventionally, one may argue that these different ways of thinking are at best complementary. From this point of view, the use of computers in historical research for mechanical tasks, such as counting, ordering, and records management, is unproblematic. However, any use of computing beyond this—and thus the whole field of digital/computational history—raises important questions, if we consider HT and CT to be in opposition. In this talk, I will look at HT and CT from epistemological and methodological perspectives and explore ways to foster a productive dialogue.“
08:00 p.m. Welcome Reception

WEDNESDAY, SEPTEMBER 30
09:45 a.m. Workshop Meet and Greet 4 Hermes PIs – open for interested Master-Students (Host: Ruth Reiche, Johanna Konstanciak)
Guided Tour through Herder Institutes Departments (I)
11:15 a.m. Coffee break
11:30 a.m. Open Colloquium: Data Sovereignty (Speaker: Jonatan Steller)
12:45 p.m. Lunch
01:30 p.m. WORKSHOP 3:
Same Sources, Different Structures: How Digital Modeling Choices Interact with Historical Questions
(Host: Dinara Gagarina)
03:00 p.m. Coffee break
03:30 p.m. WORKSHOP 4:
Don’t trust your authorities – Epistemological and practical challenges in working with multidisciplinary authority data and complex controlled vocabularies (Host: Felix Köther)

THURSDAY, OCTOBER 1
09:45 a.m. WORKSHOP 5:
I do not dare to decide! - Modelling and Processing Vagueness and Uncertainty in Historical Texts (Host: Cristina Vertan)
11:15 a.m. Coffee break
11:45 a.m. Guided Tour through Herder Institutes Departments (II)
12:45 p.m. Lunch
01:30 p.m. Data Carpentry 2: „Quarto“ Part I (Host: Steffen Uhl and Robert Zwick)
02:45 p.m. Coffee break
03:00 p.m. Data Carpentry 2: „Quarto“ Part II (Host: Steffen Uhl and Robert Zwick)
04:30 p.m. Guided City Tour
07:00 p.m. Dinner

FRIDAY, OCTOBER 2
09:45 a.m. WORKSHOP 6:
Human-Centered Evaluation: Studying users, workflows and digital tools (Host: Christin Kreutz)
11:15 a.m. Coffee break
11:45 a.m. Podium Discussion
12:45 p.m. Closing Remarks
01:00 p.m. Lunch