UZH Open Science Projekte 2025-2026

Advanced Research Data Environment for the Arts and Humanities

Lead: Prof. Dr. Tristan Weddigen (Kunsthistorisches Institut)

Open Research Data (ORD) has become essential for advancing research in the Humanities and for developing sustainable research infrastructures. The Swiss Art Research Infrastructure (SARI) and its partners have played a key role in establishing best practices, ontological frameworks, and strategies for integrating Machine Learning (ML) into cultural heritage data. This project builds on these foundations to enhance the visibility, accessibility, and adoption of ORD standards, particularly by non-academic institutions in the GLAM sector, which hold the majority of source data for Digital Humanities. It also aims to strengthen international outreach by engaging with European research networks and infrastructures.
The project aims at expanding community involvement through training events and specific educational materials. Moreover, SARI will make its Semantic Reference Data Modelling framework (SRDM) more accessible thanks to an improved, intuitive interface designed for non-expert users. SARI will also enhance interoperability with important national and international data infrastructures and stakeholders, namely DaSCH and a potential future EOSC.ch node. Another key objective is to establish a more substantial ontological basis for documenting the provenance of AI-generated data, making ML classifications more transparent and reusable. The project explores the use of Large Language Models (LLMs) to automate metadata generation and create user-friendly textual descriptions that facilitate understanding and reuse of provenance metadata.
Led by the University of Zurich / SARI in collaboration with ETH Zurich and DaSCH, this initiative will drive ORD adoption beyond academia and foster interdisciplinary innovation and international collaboration.

Towards FAIR Practices in Bioimaging

Lead: Virginie Uhlmann, PhD (BioVisionCenter UZH)

Modern microscopes capture incredibly detailed images of living systems at all scales, from single molecules to entire organisms. These powerful instruments generate massive amounts of digital image data that could revolutionize our understanding of life - but only if scientists manage to effi-ciently analyze them. Currently, most of this valuable data remains underutilized because each microscope produces its own kind of images, making it hard to manage all the data together and to create analysis tools that work across different types of microscopes. To address this, the international scientific community has been developing open standards for microscopy data (referred to as bioimages), but Switzerland's involvement in these efforts remains limited despite the country being a leader in microscopy technologies.
The BioFAIR project aims to help Swiss scientists adopt the community-defined open standards for bioimaging data in two ways: by creating free and reusable tools to support the management and analysis of microscopy images in the open file format adopted by the community, and by organizing meetings where researchers are invited to contribute to shaping the future of open bioimaging data. Led by the recently-created BioVisionCenter at the University of Zurich, the project will bring together experimentalist and computational researchers working with bioimages to ensure it responds to practical needs. The BioFAIR project will create the resources and promote the interactions needed to bring microscopy data into the open science era, positioning Switzerland to play a key role in defining international open practices for managing and analysing bioimages.

Data Stewardship at the University of Zurich (UZH)

Lead: Dr. Susanna Weber  (Universitätsbibliothek)

In the initial project, the University of Zurich (UZH) established a university-wide Data Stewards Network to provide domain-specific support for FAIR research data management. The network consists of 20 UZH researchers and technical staff, along with 8 swissuniversities co-financed data stewards located at technology platforms. These data stewards serve as bridges between researchers and university support services, helping implement FAIR data practices and providing guidance on research data management.

The next project phase will expand this initiative in three ways. First, it will broaden coverage by adding two new ORD specialists: one at the Center for Reproducible Science and another at the Department of Psychology. This addresses the current limitation of support being primarily concentrated at technology platforms. Second, it will enhance support for the network's 20 unfunded data champions through the creation of a knowledge base and improved onboarding and training programs. Third, it will increase awareness of data stewardship activities to engage more UZH researchers and technical staff, building a more robust and inclusive network of expertise and allowing more researchers to benefit from support.

The project focuses on long-term sustainability by integrating data stewardship into UZH's existing organizational and financial structures. The university has demonstrated its commitment through permanent funding for the network coordinator position and temporary funding for technology platform-based data stewards. Future efforts will continue to focus on embedding data stewardship capabilities within existing university financial and organizational structures.

Database on Intergovernmental Cooperation in Education

Lead: Dr. phil. Patrick Montjouridès (Institut für Erziehungswissenschaft)

The DICED project aims to create a global resource for scholars, students, citizens and stakeholders engaged with intergovernmental cooperation in education. The growing number of international actors, legal instruments, and cooperative instances has made it increasingly difficult to form a com-prehensive understanding of international education resolutions and recommendations. SDG 4, central to the Sustainable Development Goals, under-scores the importance of advancing global education and fostering participatory, transparent dialogue. Global governance of education is influenced
by diverse actors and mechanisms, often extending beyond state-centered perspectives. Nevertheless, international resolutions, agreements, and declarations remain pivotal as they provide key reference points for global education governance. Examples include the 2030 Agenda for Sustainable Development, UNESCO's International Conference on Education (until 2008), and the Council of Europe's Standing Conference of European Ministers of Education (since 1959). However, significant barriers hinder systematic exploration of intergovernmental cooperation in education. Resolutions
are often dispersed across organizational websites or archives, with limited institutional memory of the debates preceding their adoption. Accessibility issues further complicate understanding the perspectives, ideas, and legal frameworks shaping global education values. DICED leverages digital humanities and computational approaches to address these challenges. By utilizing standards like the Text Encoding Initiative and UNESCO-OECD-Eurostat Data Structure Definitions, the project will create a structured, accessible database and accompanying standards and tools for querying,
reading and processing data. This resource will enable the global education community to systematically and diachronically explore international resolutions and recommendations, enhancing replicability, verifiability, and understanding of global education governance studies.

ORD Practices for Language Data across Disciplines

Lead: Dr. Cristina Grisot (Linguistic Research Infrastructure)

This one-year project aims to foster collaborations and support researchers in making better use of language data in their work. Language data is essential for understanding communication patterns, patient-doctor conversations, historical texts, political discourse, and much more. However, working with this type of data can be complex, especially for researchers in fields like medicine, history and political sciences, theology, financial and business communication, translation, terminology and interpretation, and language education.
The project will begin by identifying the skills and knowledge that researchers need to manage language data effectively and in compliance with the Open Research Data paradigm in these diverse fields. Next, we will build a research community composed of researchers, including PhD students, early-career scientists, and data stewards, who will participate in the formulation of recommendations for user need-driven training sessions necessary to acquire solid research data management knowledge. Finally, based on these recommendations, we will create a compendium of best practices with training sessions designed to fill gaps in knowledge and provide practical guidance.
By the end of the project, we aim to make working with language data more efficient and to promote collaboration between researchers from different disciplines. This initiative will help advance Open Research Data practices, ensuring that scientific discoveries are more transparent, reusable, and beneficial for everyone.

Unlocking Collections Data

Lead: Dr. Alice Hertzog (Völkerkundemuseum)

The Ethnographic Museum at the University of Zurich holds around 57’000 objects from Asia, the Americas, Oceania and Africa. As such it plays a vital role in preserving, interpreting and sharing the cultural heritage of diverse communities worldwide.
Exhibitions offer the public a chance to engage with objects from our collection. Our doors are always open, and we strive to make our museum ac-cessible to everyone. Now, we’re seeking to expand that accessibility by making our collection data available through an online database.
Sharing ethnographic collections online can enhance accessibility, foster inclusion, and promote global collaboration. However, it also raises significant ethical concerns. Many objects hold sacred or confidential significance for specific communities, and unrestricted online access may lead to misinterpretation, exploitation, or cultural harm.
Respecting Indigenous knowledge and rights is key. This includes adhering to cultural protocols and responding to community requests regarding the appropriate sharing of their heritage. Additionally, questions of consent must be carefully considered, particularly given that many objects were acquired in colonial contexts.
This project aims to identify categories of objects and associated data that require careful handling due to their cultural sensitivity. By collaborating with researchers and community partners, we will develop ethical solutions that balance the benefits of digital access with respect for the cultures and communities from which these collections originate.

Introducing ORD Best Practices for Eye-Tracking Data

Lead: Prof. Dr. Lena Jäger (Institut für Computerlinguistik)

Humans rely on their vision in almost every scenario, making eye movement data valuable across diverse fields such as psychology, linguistics, eco-nomics, human-computer interaction, and machine learning. However, a major challenge of eye movement research is the fragmentation of da-tasets—many are not publicly available, scarce, domain-specific, and lack interoperability. To address these issues, we propose the development of global open research data (ORD) standards for eye movement datasets to enhance accessibility, reusability, and cross-disciplinary collaboration. To support researchers in applying these standards, we are presenting a pipeline that can be used to work with and create eye movement data in a FAIR way. By establishing common guidelines and implementing the necessary tools for applying these guidelines to eye-tracking data collection, format-ting, and sharing, we aim to break down existing barriers and foster innovation across research communities.

An Interactive Exploration Infrastructure for Digital Scholarly Editions

Lead: Prof. Dr. Jürgen Bernard  (Department of Informatics)

Open Research Data (ORD) promotes transparency, reproducibility, and knowledge sharing across disciplines. In the digital humanities, digital edi-tions provide access to annotated historical, literary, and legal texts. However, many heterogeneous editions remain isolated, limiting their potential as interconnected research resources. To increase awareness and reuse of editions and pioneering research infrastructures (like hallerNet or correspSearch), a visual interface is needed that allows diverse stakeholders to interactively engage with editions, explore their characteristics, and un-cover edition connections.
In ORD Xplore 1.0, we developed a promising prototype that bridged gaps between heterogeneous digital editions using interactive techniques, initially based on 6 UZH editions. We now face the ambitious challenge of integrating 60 heterogeneous editions in ORD-Xplore 2.0— a scalable web interface allowing different user groups to explore, compare, and analyze digital editions within a unified ORD infrastructure. The approach combines advanced data science, human-centered AI, and interactive visual analytics methods with user-centered design, supporting digital librarians, edition curators, researchers, and the broader academic community. Users can efficiently discover relevant editions, edition characteristics, and edition connections. Interactive structural comparisons of annotation practices, multiple edition perspectives, and feedback mechanisms will enable different stakeholders to contribute exploration insights, edition knowledge, and data corrections, enhancing collaboration, beyond traditional search and filtering.
By bridging 60 isolated digital editions, this project promotes ORD standards in three case study domains new to ORD-Xplore, enhances interdisciplinary research, and strengthens the digital edition landscape in Switzerland and beyond, ultimately creating a sustainable, community-driven ORD interface for long-term engagement and scholarly collaboration.

A SECure Framework for Data Sensitivity Assessment and Access Control in the HUManities

Lead: Prof. Dr. Botakoz Kassymbekova (Historisches Seminar)

Research data collected from participants in regions experiencing geopolitical turmoil presents significant ethical challenges, as consent for publication is often obtained, yet concerns about participant safety and ethical responsibility persist in the absence of established best practices. This project suggests addressing the challenges of storing and sharing sensitive data by drawing on an ongoing humanities research initiative that involves inter-view recordings from individuals in Eastern European countries who discuss their experiences of physical and sexual violence during violent conflicts. Here, a secure technical infrastructure, clear ethical and legal guidelines, and rigorous access control measures are essential to protect the interview-ees. The SEC-HUM project aims to serve as a proof-of-concept for safe storage solutions in small research projects. We intend to implement a secure technical solution while addressing in-depth the legal, ethical, and technical questions that arise in managing sensitive data. Based on this process, we will develop guidelines, questionnaires, workflows, and documentation to standardize processes and provide future research projects with a comprehensive framework for sharing highly sensitive research outputs. This framework will ensure that research data remains findable and, where possible, accessible while complying with legal and ethical standards and implement a secure infrastructure for sharing research data when possible. We will leverage our networks to disseminate SEC-HUMs outputs at institutional and national levels, guaranteeing a long-lasting impact in the ORD landscape.

SoundtrackID

Lead: PD Dr. Simon Spiegel (Seminar für Filmwissenschaft)

The SoundtrackID project addresses a critical gap in audiovisual (AV) research by providing a semi-automated tool to identify film music and assess its copyright status, thus enabling researchers to determine whether films are in the public domain. The project tackles a key challenge: while films may seem easily accessible in the digital age, legal and technical barriers severely restrict their use for research. A central issue is that copyright laws often prevent films from entering the public domain, because the music used is not yet public domain. Using acoustic fingerprinting technology – already employed in apps like Shazam – SoundtrackID will identify the music within films, cross-referencing it with databases to determine its copy-right status. The pipeline involves uploading films, extracting metadata, identifying music passages, and consulting various APIs to retrieve composer
details. This approach is novel in the context of copyright clarification for films and will significantly reduce the time and effort required for legal assessment. Collaborating with international partners in media studies and film archives, the project team aims to integrate the tool into a broader AV metadata infrastructure. By doing so, SoundtrackID will promote ORD principles in film and media studies, facilitating wider access to AV heritage for researchers and the public.

Fostering bacterial genomic data FAIRification by extending IMMense and SPSP

Lead: Dr. Tim Roloff Handschin (Institut für Medizinische Mikrobiologie)

The Swiss Pathogen Surveillance Platform (SPSP) is an infrastructure to store and share genomic data from microbes and is seen as positive example across Europe. Labs, data scientists, authorities and federal offices heavily use SPSP since the COVID-19 pandemic, especially for viruses. The objective of this project is to extend SPSP for bacterial genome sequences. IMMense is the analysis pipeline processing bacterial sequencing data within SPSP and is being developed as an open-source data analysis pipeline at the Institute for Medical Microbiology at UZH. With a steadily growing number of bacterial species being added to SPSP, new sequencing data types becoming relevant and species-specific analyses being requested, IMMense has to be extended and partly redesigned to fulfil all the needs. Reproducibility of analyses within such a clinically relevant pipeline is critical. Therefore, all updates and feature additions require careful quality control (QC) and validation. For this, a validated in-silico reference dataset of all relevant species could be used.
SPSP can currently store bacterial data and share it with other repositories and authorities but so far there is no front-end allowing scientists to explore the data generated by IMMense which limits SPSP’s usefulness.
In this project, we plan to extend IMMense to cover the current top priority analyses needed for SPSP, generate a reference dataset to regularly benchmark the quality of the output of IMMense, and develop a front-end for bacteria for SPSP to visualize the results generated by IMMense.

New Methods for Accessing Visual Open Research Data in the Digital Humanties

Lead: Dr. Darío Negueruela del Castillo (Kunsthistorisches Institut)

Understanding art and historical images requires more than just identifying what's in a picture - it needs deep knowledge of history, culture, and con-text. While artificial intelligence (AI) has become incredibly powerful at recognizing objects in images, we don't yet have reliable ways to test if these AI systems can truly understand images the way humanities scholars do. This gap makes it difficult to develop better AI tools for humanities research and holds back progress in the field.
We tackle this problem by creating a comprehensive set of tests (called benchmarks) to evaluate how well AI systems can handle complex tasks in studying art and visual culture. These tests will go beyond simple image recognition to assess if AI can understand historical context, make meaningful connections, and provide insights that are actually useful for humanities research.
The project has three main goals. First, we will develop these much-needed testing methods to understand what current AI systems can and cannot do. Second, we will create improved AI models specifically designed for analyzing art and historical images. Third, we will build user-friendly tools that researchers can use without needing technical expertise.
By making these resources openly available, the project aims to bridge the gap between AI technology and humanities research. This will help scholars from different fields work together more effectively and unlock new possibilities for using AI to study and understand our visual cultural heritage.

Creating a Sustainable Funding Scheme for Diamond Open Access through Institutional Collaboration in Switzerland (CoDOA-CH)

Lead: Consortium of Swiss Academic Libraries
Partner an der UZH: Margit Dellatorre (Universitätsbibliothek) und Samuel Nussbaum  (Zentralbibliothek)

Diamond Open Access CH (DOACH)

Lead: Dr. Andrea Hacker undDr. Elio Pellin (Universitätsbibliothek Bern)
Partner an der UZH: Margit Dellatorre  (Universitätsbibliothek) und Samuel Nussbaum  (Zentralbibliothek)

National Approach for Interoperable Repositories and Findable Research Results (NAIF)

Lead: Dr. David Johann und Simon Willemin  (ETH Bibliothek / ETH Zürich)
Partner an der UZH: Dr. Andrea Malits (Universitätsbibliothek) und Stefan Vogt  (Zentrale Informatik)
Projektseite

Open Online Course on Good Research Practices (OOC-GRP)

Lead: FernUni Schweiz
Partner an der UZH: Dr. Fabio Molo  (Center for Reproducible Science and Research Synthesis)

Report on Open Access in Law (ROAL)

Lead: Dr. Simone Rosenkranz und Dr. Dominik Matter (Universität Luzern/ZHB Luzern)
Partner an der UZH: Barbara Berchtold undMargit Dellatorre (Universitätsbibliothek)
Projektseite

Re-Use Standards for Editions Data with Application Programming Interfaces (ReSED API)

Lead: Dr. Elena Chestnova (Università della Svizzera italiana)
Partner an der UZH: Dr. Yann Stricker (Universitätsbibliothek)
Projektseite

Semantic TEI 2: Semantic text publishing as open research data for a broader audience (Sem TEI 2)

Lead: Dr. Elena Chestnova (Università della Svizzera italiana)
Partner an der UZH: Swiss Art Research Infrastructure (SARI)
Projektseite

Swiss EOSC Node Prototype (SENPro)

Lead: Prof. Dr. Thomas Schulthess  (CSCS / ETH Zürich)
Partner an der UZH: Dr. Andrea Malits (Universität Zürich)