Visions, needs and requirements for (future) research environments: An exploration with ERC Grantee Toma Susi
by: Barbara Sanchéz and Katharina Flicker (TU Wien)
Researchers are at the very heart of the EOSC: So what do researchers really need to do cutting-edge research? How do they think the EOSC could support them in their endeavors? Let's see what ERC grantee and physicist Toma Susi has to say.
An exploration with Toma Susi, Assistant Professor for Physics of Nanostructured Materials, University of Vienna; holds an ERC Starting Grant;
Toma Susi is active in Open Science policy and advocacy as Vice-Chair of the Young Academy of Europe (YAE); check out his website Mostly Physics https://www.mostlyphysics.net/
Data Sharing in the interest of reproducibility and re-usability
KF: What is your work currently focused on?
TS: My work focuses on material science and more specifically electron microscopy. What we do is we look at the atomic structure of different materials, and their defects and dynamics and things like that. In addition to the experimental work, we do quite a bit of modelling. We use methods based on quantum mechanics to model the properties of these materials and then try to understand what is going on by combining the experiments with the theory and correlating the two.
KF: Do you have data challenges that EOSC could help you with?
TS: Yes. We generate two types of data. We generate electron microscopy images, which are in principle normal image files except that they are 32 bit. If you open such a file on your normal computer without using a specific program, you won’t see much because the contrast range in such images is much wider. And if there was a way to have an open data repository where we could just upload a bunch of images and the website would make preview images that you could just browse in your normal web browser. And you could of course also download the original data files. That would be very useful. The other challenge for making data available are the simulations we run to model the properties of materials. These generates files that are pretty big and are typically not shared at all between researchers. We generate such files to the tune of 1 to 2 TB per year. We keep them for the time until the publication is accepted. Maybe a little bit beyond that. Then we have to delete them, we normally cannot store long-term such big files.
KF: What would you need from the EOSC in order to resolve the challenges you face? Which services would help you most?
TS: I do not currently see us doing much of this work we do in the cloud. We can do the analysis locally and it probably is easier. For us it is about open science. That is where the cloud could be very useful. And of course, not just have static files sitting on a server, but something that you and other people could access and re-use in various ways.
KF: What is the main added value that you see in the EOSC?
TS: Having one place for all European research data that is openly available on the web. I think that is the valuable and worthwhile goal.
KF: Do you believe that the EOSC can boost interdisciplinary research in Europe?
TS: I don’t think that interdisciplinary research really is a data problem. That is more of a knowledge and communication and networking problem and of people actually finding these collaborations. You cannot just take a dataset that you don’t understand and do research with that. You have to understand the data and to do that you have to cross boundaries and need somebody from the other side helping you and instructing you. Even good metadata probably isn’t enough. You can describe the data. But unless you understand the discipline specific meanings of those descriptions, it doesn’t get you very far. In terms of helping within the discipline of nanoscience and physics – if more of the data that we generate or that people in the field generate was more openly available and easily accessible and perhaps even some analysis that we could run in the cloud, we would use that. I think it would improve at least our ability of verifying results. But the question is then at which point of the research cycle this is going to be incorporated in. Is it post-publication, or even part of the peer-review process? Or Is the data already part of the submission? Or even before the submission like in a pure open science approach?
KF: Do you have preferences when it comes to when to share the datasets?
TS: That also depends on how easy it is to share the data. Currently you have to do most of the work manually. If you could just drag one zipped file to a cloud and everything else is handled somehow automatically, it would be much easier to share the data in the beginning of the peer-review process. Sharing something before that is another story. There you do get into –I think legitimate – fears of being scooped, if you share your latest findings months before you have a manuscript ready or before you submitted something. I do think that we would be hesitant with that. Because we can collect very specific types of data, and just finding the right samples, making the appropriate measurements, getting a good dataset – that is a lot of work and if you just share it before you are ready to publish, perhaps then somebody will take the data and will do the analysis that you wanted to do. That accelerates science. But it doesn’t accelerate our careers in any way. So I think that at submission we would be comfortable with sharing data.
KF: In your domain, which challenges can you or will you address with the EOSC, or with more open data available which would be impossible to tackle alone?
TS: It would be valuable to have a way of fulfilling open data mandates in a user-easy and painless way for putting our data out there, with our papers and having it potentially be cited and re-used. That for us is the biggest motivation for doing open data. Currently we do some data open, but it feels that you are just uploading stuff on the cloud or on a server and that is it. Nobody ever uses it. Nobody ever cites it. It is nice to be open, but if there is no tangible benefit of it to anybody, then it feels a little bit like a waste of effort.
KF: What is the meaning of open science and how does it increase the value of your work as a researcher?
TS: For me it has to do with reliability, verifiability, and the re-use of the research we do. Currently at least it is tied to publication. Every single part of the publication needs to be open. The paper itself needs to be open. The data needs to be open. Anybody should be able to reproduce the analysis and whatever steps are necessary to reach the conclusions that are in the paper. So they should have the data available, and all the analysis available and they should be able to reproduce this whole chain and reach the same conclusions. That will help to improve the reproducibility of research.
KF: How does the EOSC help to drive open science and facilitate open access?
TS: Making it very easy to various types of different data in a way that is helpful and useful.
KF: What are you missing in the discussion? What do I need to tell everyone?
TS: There needs to be a very clear explanation of what the EOSC is. Why are we doing this? How is it going to work in practice? There needs to be a single page with the full story. Not tens of pages of interconnected explanations. One webpage where I can go to understand what is going on and why I should care. There needs to be a clear message about what it is and why we are doing this. In terms of adoption, it needs to be simple and painless. If we have to jump to hoops or if fill in pages of metadata manually it is not going to be widely adopted.
TU Wien is partner in the H2020 project EOSCsecretariat.eu, delivering 360° support to the EOSC Governance. Specifically, it addresses the need for the set-up of an operational framework supporting the overall Governance of the EOSC.
EOSCsecretariat.eu contacts at TU Wien:
- Andreas Rauber (PI)
- Barbara Sanchéz
- Paolo Budroni
- Katharina Flicker
- Juliana De Mello Castro Giroletti
- Bernd Saurugger