Data management plans¶
Data management plans are a catchphrase these days, mainly because funders are requiring them now. This is for a good reason - researchers often focus on their papers, and making good use of the data gets forgotten. Funders pay a lot for research, and they want all the possible value for society.
However, it is worth doing a bit of planning about data, even aside from the required bureaucratic exercise. It is true that researches focus on the next paper. Data has long-term value even inside Aalto, and if you don’t try hard it will get lost.
In this section, we outline recommend ways to use Aalto resources for different use cases.
No matter what your project, you want to start by thinking how you will handle your data (this can be “real data”, notes, code, papers, etc). This will make sure that your team works together well and doesn’t end up with a big mess in a few months - or that you can’t work together because you can’t share information. For this, see the A4 DMP template. This site is focused on practical DMPs.
Suggested DMP for large experimental data (TODO)
Suggested DMP for simulations or computer-generated data (TODO)
Suggested DMP for data from humans (surveys, interviews, etc) (TODO)
There are plenty of other good resources about making funder DMPs.
At Aalto, the RIS grantwriters have taken responsibility for helping to make good funder DMPs.
The Aalto RDM pages have a subsection dedicated to data management plans.
The DMPTuuli is a combination template, instructions, and web form which makes it easy to do the mechanical assembly of DMPs. They also have public docx/pdf templates which can be used even without the web form. Aalto recommends this service, though be aware it helps you fill out a form, not plan your work.
As some concrete suggestions:
Funders are especially concerned about sharing, preservation, reproducibility, and dissemination but probably can’t evaluate too much about the practical side of things.
You can mention that you will follow the Aalto RDM policy, which covers mainly opening and licensing. The policy still allows you to make your own choices, but it sounds quite good if you refer to it and say you will follow it.
For data storage considerations, you can say that your department/Science-IT provides data storage services (for Science-IT departments) and has a data storage policy which you will follow: citation and/or full text.
Help! I need a DMP right now!¶
If you are reading this, you probably have a grant deadline and you need to do something right now. Use the resources above, but here is some more advice:
Read the data management outline on this site. You should be able to pull many of the practical pieces (storage, confidentiality, archiving, etc) from here. Read this first!.
Read the Aalto-level guidelines. These are quite abstract and high level, and might tell you what people think is important but not tell you how to do stuff.
To internally organize things, you could start with the A4 DMP template. This can’t be used for something you submit, but lets you know the big picture. If you fill this out first and give it to someone, they can guide you in making the next version.
Use the DMP Tuuli tool to prepare the DMP. It just makes a final document you can download (you could do the same using a word processor), but breaks everything down into a nice form.
If you don’t like the idea of a web form, the templates seem to be available publically, too. These seem to have roughly the info as the DMPTuuli web forms.
Why do they want DMPs? What should it include? Answering these will help you to know what to write, since there is not near enough room to make a plan that contains everything you need to know personally.:
The main purpose is to make sure that other researchers can use your data as easily as they can use your published papers. Can other researchers access your data? Can your results be reproduced?
Most likely, whoever is reading doesn’t care that much about the actual day to day data storage and so on, but more of the big picture: licensing, opening, archiving, sharing, preserving, expanding, securing.
If you produce your own data, how can others use it? Funders want open, but by giving good justification you can do whatever you need. If the data comes from others, then can you re-distribute (even for validation) or would others need to request it from the source?
How software you make related to data processing (and really all software) will be handled. Even if data can’t be released, software can be open sourced which allows reproduction of results and some sort of validation.
How you preserve data for future use: both for you, and for others. This is especially important. Also, how will data be understandable in 50 years? Is the program that will read it gone? Do you have a README? Is your data in a field-specific standard structured format? Is it opened and does it go into an archive which will be around in 50-100 years (anything managed by you or Aalto specific isn’t a credible option for this)?
You should mention how you will follow the “Aalto Research Data Management Policy and related guidance”. The policy just says “you will make strategic decisions”, so sounds good to the funder while not binding you to anything.
For storage, organization, confidentiality, etc, you can say you will follow the Science-IT data management policy. This isn’t requirements for you, but the default services we offer for data storage (designed to keep data safe and secure, and uuushareable). It also sounds good to say. (see the outline)
Model Academy of Finland DMP¶
You can see the Academy’s detailed info in their supplement. This guide isn’t to replace their guidelines (there is a lot there that isn’t duplicated here), but make it clear what the Aalto correspondences are. You can also see the Aalto guidelines, but this is also a bit abstract to be immediately usable.
With all the time spent on writing your plan, don’t forget to do something useful, too.
General description of the data
No specific extra advice here - see academy guidelines.
Ethical and legal compliance
For identifiable human data, say that you will follow the Aalto personal data policy. In particular, data will only be stored only on systems meeting the Aalto guidelines for personal data storage. Preferable, store this on the department network drives only - not on personal computers. You can request ethical evaluation from the Aalto Research Ethics Committee. Is Finland, this is required in quite few cases, but publishers are requiring this more and more often. Thus, you may want to check your journal requirements and request ethical evaluation anyway.
Data always will be made available under the Aalto data management policy. (You can commit to this, because the policy only says you should make decisions “strategically” so there are actually no obligations.)
Software will be made open source if it matches the criteria under the Aalto open source policy. If software exceeds that criteria, there will be discussions with Aalto innovation services for commercialization or licensing.
There are plenty of other intellectual property concerns which I can’t go into here, and you need to study yourself. Aalto Research and Innovation Services has lawyers which can help with this - you can consult in advance or say you will use them.
Documentation and metadata
It is harder to comment on this because it is so field-specific. Make sure you have READMEs and documents.
Everyone talks about “metadata” but this is such a broad term that it is essentially meaningless. I personally put this into three types:
Cataloging: You can say that the metadata required by your repository will be used.
Necessary to understand: you will use README files, use formats that are self-describing such as CSV files with useful headers and comments, include code, and whatever is needed to make someone understand the data later (including yourself).
Necessary to automatically process: data should be automatically usable with the least amount of manual effort. This is highly domain-specific, and depends on if your domain already has standards to make this possible. Use the best possible practices here, taking into account cost vs benefit.
Storage and backup during the research project.
Aalto really excels here. Basically, just use the Aalto network drives. This storage is large, free, shareable, snapshotted, backed up to an offsite datacenter. Access is controlled via Aalto accounts plus unix groups. If people need to make other copies (and it’s allowed for security reasons), they can. Big data is stored on Triton cluster from which it has direct access to any computational power you may need.
Opening, publishing, and archiving the data after the research project.
This gets more abstract, and really depends on what you want. There are many options, and maybe it is best to consult the Aalto page on this, though it’s again rather abstract.
You can check the services page to see what common services are available. If you don’t have any more specialized repository to use, Zenodo is a good choice. Always prefer a specialized, domain-specific repository if you can. Don’t say it is archived on Aalto resources, since you or Aalto can’t commit to hosting things or the long term.
You can say that organization of data is a part of research, though the extra requirements needed to open are small. Give some estimate of the total/extra amount of work needed.