Barcamp Open Science 2021: Opening up New Perspectives

Privacy preserving Open Data

by Antonella Succurro

The Horizon2020 guidelines on FAIR (Findable, Accessible, Interoperable and Reusable) data (PDF) states that data should be “As open as possible, as closed as necessary”. When it comes to sensitive data it is then up to the individual researchers or institutes to decide how to “open” them: they can opt for a very conservative approach (not share at all) or develop creative and innovative ways to share sensitive information. Common privacy preserving data sharing practices include data anonymisation and data aggregation, but both approaches can present challenges. In the first case, there is a risk of de-anonymisation through, for example, cross-referencing with additional data. In the second case, the more aggregated statistics are gathered, the higher the chance that some information gets identified.

This session highlighted how the support from data privacy officers or external trustees is often not sufficient to face the challenges of highly specific data, which might come from new technologies and have yet to be standardized. For example, genomics data are considered impossible to anonymise without losing the biological value for research purposes. We hence need a coordinated effort to build infrastructures providing training and services for harmonised data sharing “best practices” within the legal framework of the General Data Protection Regulation (GDPR).

Session Pad

Open Knowledge Maps Custom Services

by Peter Kraker and Michela Vignolio

Today we are facing a discoverability crisis that makes it hard to stay on top of the latest research. Open Knowledge Maps (OKMaps) is an open visual discovery tool based on knowledge maps that supports researchers, students and practitioners searching for relevant scientific knowledge. As a community-driven organisation, OKMaps involves the broader user community from early on in the development phase.

In this spirit, the OKMaps team invited the Barcamp Open Science participants to discuss early ideas for the brand new ‘custom services’ of OKMaps. The customisable cloud services will be embedded in libraries’ discovery services to add instant visual capabilities. The session participants reacted positively to this concept and suggested additional interesting use cases. One was to integrate knowledge maps with institutional researcher profiles to add a visual presentation of authored publications. Another proposition was to use knowledge maps to display reading suggestions for students to have a more clearly arranged alternative to confusingly long lists.

The input provided by the community helps the OKMaps team immensely to shape the custom services design according to user needs. The custom services are funded through the Open Knowledge Maps supporting membership programme. Supporting member organisations are directly involved in the decision-making process of Open Knowledge Maps. More information on the programme can be found on the OKMaps website. Please contact us if you want to get involved via email to info (at) openknowledgemaps.org or via Twitter @OK_Maps

Session Pad

Knowledge Inequity & Open Science criteria as a way out of it?

by Felicitas Kruschick

Is Open Science automatically leading to Knowledge Equity? We discussed this question and found out that it is not that easy. Everyone needs to constantly reflect on the power dynamics in which everyone has to analyse and discuss their project and work (which was considered not that easy due to time e.g.). Who is benefitting how from projects and who is not? Who is leading through theoretical discourses and why? Whose voices are being heard and whose are staying silent and why? Everyone needs to ask themselves: Who is left behind and in which way? Only if everyone continuously reflects on those questions leading to different answers for different contexts and realities everyone is able to face the challenge of Knowledge Equity.

The question of addressing must never be forgotten. Who is addressing whom in which context? Who is meant by ‘we’ and who by ‘the others’? If one thinks of Said’s theory of Orientalism (2003), it becomes obvious that the question of knowledge inequity can only be posed in the context of power, which has direct consequences for the addressing of the ‘other’. Whether in the context of science and research, politics or economics, internationally or locally – when a ‘we’ is assumed, an ‘other’ is manifested, which can have the effect of reinforcing injustice at the manifest level. One of the main outcomes of the session was therefore that one should never stop asking oneself what position one holds in which context and who is being downgraded at the same time. But what emerged as equally undoubted was that the question of knowledge injustice must be seen as a dilemma, as it can never be fully achieved.

A second important point was the question of why science and research do not ‘work’ more with Open Science criteria. It was noted that science is perceived as being stuck in ‘old’ infrastructures. It takes effort and courage to break down these old beliefs, which in our opinion are based on financial aspects and reputation, but also a great deal of uncertainty about the maintainability of scientific quality, and to introduce new ideas. We have also found out that Open Science affinity probably differs from department to department. In social science disciplines, the use of Open Science is less established, as the argument of sensitive data is often used here.

Session Pad

Further Readings:

• A Global Ethics Code to fight ‘ethics dumping’ in research
• African Digital Research Repositories
• Decolonizing Scholarly Communications through Bibliodiversity
• Said, E. (2003): Orientalism. Western Conceptions of the Orient. Penguin Books.

(Re-)using available research data in the social, educational, behavioural and economic sciences

by Thomas Lösch and Sonja Bayer

(Re-)using available research data means that research data provided for example by research data centers is used in a way that was not necessarily intended while the data was being collected. (Re-)using data has many advantages as it realises the potential of Open/FAIR data by making good use of available resources.

The session was intended to get an overview of what the current state is. Is data used? What works, what doesn’t and why?

We mainly identified barriers and some enablers for (re-)using research data:

• Research data needs to be well-documented so (re-)users can see how the data was created – especially with regard to nuances in different labs or communities (for example, slightly different ways to measure the same thing). This is even more relevant for an interdisciplinary (re-)use of available data.
• A centralised point of access to research data helps researchers a lot in finding useful data – for example, this is the case in astrophysics.
• Policies of funders as well as application procedures should allow the (re-)use of data as a standard possibility.
• Moreover, funders can foster data reuse by including it in their materials. For example, grant application forms could explicitly require a statement about which available data will be used (or why new data needs to be collected). Also, funders could directly point to trustworthy repositories.
• In scientific communities, (re-)using data is often not valued as much as collecting data yourself. Sometimes, it is seen as more innovative to collect data in an amazing new way.

An opportunity to dive deeper into the topic might be in September 2021, where we plan to organise a barcamp dedicated to the topic of (re-)using research data.

Session Pad

How to support Open Scientists within your university?

by Philipp Zumstein

Supporting researchers with access to information or providing them with hardware and software are traditional services for libraries or IT service units within a university or other institutions. When researchers are practicing Open Science they may have further needs for support and that was the topic of this session.

As a result of some quick starting questions among the participants, we see some hints of which aspects of Open Science support are important and how much support is currently there:

Philipp Zumstein: What aspects of Open Science are important and how much support you get for them?

There are several areas of Open Science for which institutional support may be lacking. It was discussed that responsibilities within an institution (for example: for Open Access it should be the library) are crucial for establishing Open Science services. Thus, the question is also, for example: Who is responsible for supporting Open Scientists with preregistration of studies? Any strategic plans for a university should also consider the subject-specific needs from their researcher areas. Many more things were touched upon in this fruitful discussion and helpful links were given. For further details have a look at the session pad.

Session Pad

Approaches of Single Source Publishing in Research and Education | Community Networking

by Tim Boxhammer, Axel Dürkop, Florian Hagen, Albert Krewinkel, Isabella Meinecke, Klaus Thoden, and Simon Worthington

The aim of our session was to network stakeholders from the Open Science community that are actively working on or are highly interested in Single Source Publishing (SSP). SSP allows producing diverse output formats (for example: HTML, PDF, JATS/XML) from only a single source (for example: Markdown or XML). Three approaches – briefly listed below – were introduced and discussed. Newcomers and experts defined the benefits of SSP focussing on the different aspects of the research lifecycle.

In the session three projects were presented:

1. At TIB Open Science Lab, Simon Worthington and colleagues are working on semantic publishing for architecture where they try to pull together a variety of data sources and complex digital objects into a textbook publication. The research context is in the German consortium of the National Research Infrastructure for Culture (NFDI4Culture) where a consortium brings together advanced open infrastructures, for example: knowledge graphs, IIIF deep image zoom, Wikidata, and PID use in 3D point clouds and video.
   See also the GitHub repo: ADA – Semantic Publishing.
2. Klaus Thoden presented the Edition Open Access Publication Infrastructure, which uses TEI-XML as the main document format for its SSP workflows. He stressed that in SSP it is important to ensure that the source format is the richest format. That is, it will contain all the necessary mark-up and metadata. The final publication formats like PDF, EPUB or an online version are created from that source, and they can be considered lossy versions that simply serve their special purpose.
3. Axel Dürkop finally gave a short introduction to Swapfire, an approach by Hamburg University of Technology (TUHH) and Hamburg State and University Library. It is a modular framework developed within the project Modern Publishing that consists of GitLab and Open Journal Systems (OJS), supported by pandoc/pandoc-scholar (multi-format converter) and hypothes.is (annotate). The framework is an effective and highly customisable approach to produce diverse output formats from a single source.

Darstellung eines Workflows für Single-Source-Publishing mit Markdown, Gitlab, pandoc und OJS. (CC BY 4.0).

The short presentations were followed by a discussion of the participants. A wish was expressed to have more hands-on material to get started. Finally, the idea to build a community on SSP was supported.

How to participate in the future? Contact: email to oa-pub.hos(at)tuhh.de or take a look at the session pad.

Session Pad

At Your (Data) Service: How Can We Expand Research Data Management ‘Service Portfolios’

by Emma Anne Harris

There is a lot of discussion regarding increasing Research Data Management (RDM) services from the perspective of researchers. However, understanding and supporting service providers is also crucial to embedding Open and FAIR Data approaches in institutions.

The session identified some clear trends in what is and is not being offered. Participants put that advice on Data Management Plans (DMPs), general training on RDM, and archiving support were all common services being provided. Whereas legal advice and support on sensitive data for example guidance on GDPR regulations, as well as information on collaboration and repository tools for example commasoft or the RIN data repository, were identified as services that participants wish could be provided. Time, workload, personnel, and lack of institutional buy-in were all identified as obstacles to being able to offer more RDM services.

The final discussion was about how participants felt they could expand their service portfolio. It was agreed that an underlying RDM strategy would be helpful. Several participants highlighted that better connections between departments, for example library and IT, would help, although these required careful and incremental improvements. Additional funding and staff would also be of enormous practical help. Others suggested that researchers themselves might be able to offer support and expertise in creating services that expand the RDM services offered, and that indeed bottom-up or grassroots initiatives, for example openmod, are more likely to be embedded in the long-term. The results of this workshop will be fed into the FDNext project, which will create resources to assist service providers in concretely assessing and expanding their RDM services.

Session Pad

Publication infrastructures that support reproducibility

by Juliane Kant

Reproducibility and transparency can be regarded (at least in experimental research) as a hallmark of research. The ability to reproduce research results using the original data and procedure (for example: code) helps to build on existing knowledge. Despite the fact that most researchers agree that reproducibility and transparency are desirable goals and part of good research practice, efforts to enhance reproducibility are still not part of everyday research.

In this session, we discussed what researchers need to publish their results reproducibly and how infrastructures can support this process. The participants mentioned that researchers need infrastructures that are easy to use, fit seamlessly in their routine workflows, are free and ideally Open Source. Examples were long-term interactive notebooks, repositories for code, electronic lab books, and platforms that link data, code and publications. These infrastructures should be based on common standards and integrate existing tools such as GitHub or GitLab, Zenodo, and R Markdown. Moreover, the continuous availability should be guaranteed. In addition, local support for researchers was deemed necessary, for instance, with regard to data management. Finally, incentives and time are essential pre-conditions for reproducible practices. Results from this session will be integrated into a project of the European initiative Knowledge Exchange on Publishing Reproducible Research Output.

Session Pad

Open Science in Elections – Engaging Politics in Election Periods

by Stefan Skupien

2021 is “Super election year” in Germany where several elections take place at federal and state level. Since science is increasingly part of the public discussion and since the current pandemic is used to showcase the success of Open Science practices, questions arise as to whether to engage with politics to gather more support for Open Science policies at the federal and state level.

Elections could bring more attention to the issue and force parties to include Open and quality Science in their programmes. Past elections have for instance led to the science barometer as a tool to check for science policies in general. Is this replicable in 2021?

In the session, we discussed the topic and noted the availability of national recommendations for Open Science (for example Austria) and the current support at the EU and the UNESCO for implementing Open Science practices and policies. However, we were cautious of differing motivations to support Open Science and the lack of awareness and priority for science/Open Science amongst policy-makers. At national and local levels there are a variety of networks that also focus on science policy for a more Open and fair Science ecosystem. We plan to keep ourselves informed about local initiatives (for example the innObeers: monthly Open Science & Open Innovation in Science meetup for Open Science coordinators in Berlin).

Session Pad

Further Reading:

Networks operating throughout Germany that also influence politics (and advocate for higher education policy aspects)

• Bündnis Freie Bildung (German)
• Hochschulforum Digitalisierung
• Wikimedia Germany
• Chaos Computer Club
• Digital-O-Mat: an election aid for net politics
• Foundations for Open Scholarship Strategy Development