Responsibility is something everyone has to deal with every day. On one hand, responsibility means that there are organizational duties to carry out such as finishing projects in time or taking care of working resources. On the other hand, responsibility is always linked to power, as it means to control something and someone - or better expressed by a famous quote connected to the Spiderman comics: “with great power comes great responsibility”.

In the context of research data, responsibility is often thought in context of FAIR. It is somewhat clear that data needs maintenance and has to be documented in order to reach its full potential. But the research data itself is rarely discussed in context of responsibility, as research data is meant to give an objective view on world phenomena.

This year’s 13th Research Data Alliance Plenary Meeting in Philadelphia was digging deeper into these aspects and opened up an important and often overseen dimension of research data. The keynote of Dr. Julia Stoyanovich [1] was certainly a center piece in this discussion as she demonstrated that data is not as objective as its reputation. As data is taken in a specific context and represents only a section of real world phenomena, data analysis supports discrimination. Predictive analysis may contain statistical bias in the model or societal bias in the data, which result in an incorrect reflection of the world. The assumption that data science is algorithmic and therefore cannot be biased is not true, as discrimination is exhibited in the data science ecosystem and technology alone will not solve the problem. Responsibility in data science is therefore a new and important frontier. Transparency, regulations, and civic engagement are ongoing attempts to reach fairness in data driven approaches such as machine learning. However, as stressed out by the RDA representatives Lynn Yarmey, Hilary Hanahoe, and Dr. Leslie McIntosh Borrelli, it is not just data scientists that are responsible for fairness in research data. The overall theme of the plenary “With Data Comes Responsibility” already indicates that everybody is appealed to take responsibility - as the “Y” in responsibility clearly stands for “you”.

Certainly, one aspect of everyone’s responsibility in context of this plenary meeting was also to make the best out of this incredible get-together of different stakeholders. The RDA plenary meetings regularly attract hundreds of scientists, service operators, publishers, funders and industry representatives and is one of the best networking opportunities in context of research data and software. The working groups reflect current topics and initiatives in the field of research data and have impact on the research data landscape.

One example is the Software Source Code Identification WG [2] which had its first physical meeting on the RDA 13th plenary meeting. Beside datasets, software plays an important role in the research process. The recognition of the specific dynamics of software and its relevance in research are still evolving and, compared to datasets, software is still treated as third-class citizen. Yet, the awareness of software as a pillar of Open Science is increasing as credit reflecting metrics are desired as incentives for Open Data. But as source code is special, it is challenging to learn from existing systems on one hand and meet the specific dynamics of source code on the other hand. The WG tackles the lack of consensus on how to choose a suitable license, cite software, relate to industry’s best practice, and make source code FAIR. Identifiers for software are needed, but are facing challenges such as a fragmented landscape and the high complexity of source code. Sustainable solutions are desired as existing solutions and services cannot run forever on goodwill and coffee.

From a personal early career perspective, responsibility also means to step out of one’s comfort zone, reach out to the participants and present what drives you in context of research data. The Early Career program of the Research Data Alliance Europe enables individuals at the beginning of their career to participate at RDA plenary meetings. As a doctoral student in my first year, this was an incredible and important experience. The opportunity to present a poster about my research project [3] was a great chance to discuss my research approach and receive feedback from experts in the field on first results and to improve my work. The diversity and internationality of RDA’s members and contributions opened up new horizons and allowed comparisons with similar approaches. But the RDA plenary meeting was also an outstanding opportunity to network with other newcomers from around the globe and become an established member of research data community. Regarding this responsibility, surely no one left empty-handed the Philadelphia plenary meeting.

[1] = Stoyanovich, Julia (2019): TransFAT: Translating Fairness, Accountability and Transparency into Data Science Practice. URL: https://www.rd-alliance.org/sites/default/files/attachment/StoyanovichP1...
[2] = WG Software Source Code Identification - 13th RDA Plenary Meeting. URL: https://rd-alliance.org/wg-software-source-code-identification-13th-rda-...
[3] = The Mysterious Case of Research Data Reuse - or a series of unfortunate measurements. Zenodo. http://doi.org/10.5281/zenodo.2621239