Open data practices in computational social sciences

Open access to research data promises far more than transparency and participation in the research process. The better usability of the data should increase the productivity of the scientific system. Science policy thus embraces the vision of innovation and the possibility of more effective commercialization of scientific results. In science, however, opening research data is often met with concern. There is, for example, the fear of being scooped. Others see a threat for their scientific careers or want to avoid inappropriate or unethical use of scientific knowledge. The research project aims to investigate how openness is envisioned, negotiated and enacted in data practices in the computational social sciences. The focus allows us to explore openness in the context of rapid digitization and novel data forms, like social media, on the grounds of social scientific practice and daily research routines. What about the data streams we leave on the web every day? How do they feed into social research? What ethical principles do researchers follow? What data from public archives are available to the researchers, and how are they using it? How do they open their own research data for further use, or even for ones under study? To what extent can participative approaches enable openness in the research process? All these questions guide the investigation of the transformative potential of open research data in the computational social sciences. The objective is to analyse and further discuss how computational social scientists can engage reflexively with the realities they wish to create. At the crossroads of science, technology and society, this habilitation project is oriented towards approaches from the fields of "critical data studies" and science and technology studies. Based on ethnographic case studies, interviews, group discussions and data workshops, I will reflect on the potential of open data practices to increase our capacities to act or react in the context of big data, algorithmic regimes and data- driven decision making. The aim of this approach is to generate the broadest possible understanding of open research data and the related practices in the computational social sciences. The results of the project will be disseminated to discourses and debate in open science policy making as well as in research ethics governance. Furthermore, they should contribute to foster the critical discussion and uptake of openness in the computational social sciences.

Open Science has become a defining reform movement of scholarly research. Since the early 2000s, librarians, researchers, funders, and policymakers have promoted sharing publications, data, and methods to strengthen transparency, innovation, inclusion, and cooperation. Yet openness does not benefit everyone equally. Access to data, tools, infrastructures, skills, and resources remains unevenly distributed, while commercial actors seek to capture shared knowledge. This project examined these tensions through the everyday data practices of researchers in computational social science. Its central question was not simply whether research data should be shared, but under which conditions openness can be responsible, sustainable, and beneficial. For many years, research on data sharing focused on researchers' motivations, reservations, and incentives. The recurring answer was often a qualified "yes, but": researchers endorsed openness in principle, while pointing to legal, ethical, technical, epistemic, or professional limits in practice. Today, funder mandates, institutional policies, and journal requirements make open data a formal obligation. This raises a more fundamental question: not whether researchers want to share, but what can responsibly be shared, under which conditions, through which infrastructures, and for whose benefit. Across several smaller studies, the project investigated openness where data access is uncertain, contested, or newly politicised: social media research after platform access became restricted; citizen social science, with its questions of participation, consent, and interpretation; a public infrastructure for access to government data; and debates on artificial intelligence, where open research data are treated as a resource for commercial development. The work combined interviews, observation, group discussions, and desk research. A key contribution is to disentangle facets of openness that are often conflated. Access, sharing, and re-use are not the same. Access concerns who can obtain or work with data. Sharing concerns how data, methods, or materials are made available to others. Re-use concerns how data travel beyond their original contexts and are repurposed. The work showed how openness is shaped by technical, institutional, legal, and financial conditions that affect researchers' agency and determine what data can be accessed, shared, reused, stewarded, or made available. Sustainable openness requires more than publishing datasets online. It depends on maintenance, documentation, negotiation, repair, infrastructures, and institutions capable of enabling, safeguarding, or collectively governing access. This is what the project describes as the politics of openness: a contested field of practice involving decisions about power, responsibility, infrastructure, governance, and public value. Practical outcomes included Data Walks and Data Tracking exercises, contributions to Open Science and research infrastructure debates, and policy input on responsible AI governance. The broader contribution is conceptual: Open Science is not simply a movement for sharing, but an ongoing negotiation over the conditions under which scientific knowledge is produced and becomes accessible, useful, stewarded, and collectively beneficial.