Research Data Management
Prerequisites / recommended skill level
The following sections in this handbook provide useful context and complementary information to this chapter:
Table of Contents
- Research Data Management
- Prerequisites / recommended skill level
- Table of Contents
- How this will help you/ why this is useful
- Chapter content
- What Is Data?
- The Research Data Lifecycle - A Model for Data Management
- The FAIR principles and practices
- Storage and backup
- Data organisation in spreadsheets
- Documentation and metadata
- Sharing and archiving data
- Steps to share your data
- Barriers to data sharing
- Personal Stories
- What to learn next
- Further reading
Research Data Management (RDM) covers how research data can be stored, described and reused. Data here is used as a generic term to encompass all digital objects. RDM is a key part in enabling reproducible research. RDM ensures efficiency in research workflows, and also greater reach and impact, as data become FAIR (Findable, Accessible, Interoperable and Reuseable). Data should be stored in multiple locations and backed-up regularly to prevent loss or data corruption. Clearly describing data using documentation and metadata ensures that others know how to access, use and re-use your data, and also enable conditions for sharing and publishing data to be outlined.
How this will help you/ why this is useful
To be able to share data that is understandable and re-usable by third parties, research data needs to be managed properly. The FAIR principles provide guidance on how to make data discoverable and reusable. FAIR data is a key component of reproducing an analysis. However, FAIR data is not a synonym for open data or quality data. This chapter lays out good data management practice to allow you to plan your data management activities at the start of your reproducible research project.
What Is Data?
Data are all digital objects that you use and produce during your research life cycle, encompassing datasets, software, code, workflow, models, figures, tables, images, articles. Data are your research asset. In some fields it’s obvious what data means - you have observations or results of simulations. However, in other fields, particularly in Social Sciences, Humanities or Arts, you may be thinking “I don’t think I have any data”. A good way of thinking about what might be classed as data that needs to be managed is to ask yourself the questions “What is the information that I need to use and write about in my paper or book?”, “What information would I need to back up my conclusions?” and “What information is needed by a third party to understand and possibly replicate the research that I’ve done?”. This information is your data.
The Research Data Lifecycle - A Model for Data Management
Research data often follows a ‘lifecycle’ which follows the research project as it evolves; here is a video that describes it. This model provides a useful basis on which to plan for research data management, from data creation at the start of a research project, through to publishing and sharing research at the end of the project, and archiving any research data for the long-term and for future re-use once the project has ended.
The research data lifecycle involves data creation, data use, data publication and sharing, data archiving and data re-use or destruction. However, data have a longer lifespan than the research project that creates them.
The FAIR principles and practices
The FAIR guiding principles for scientific data management and stewardship (Wilkinson et al., 2016) have been developed as guidelines to improve the findability, accessibility, interoperability and re-usability of digital assets; all of which will support research reproducibility. Defined and endorsed by a growing community, these principles put a specific emphasis on enhancing the ability of machines to automatically find and use digital objects, in addition to supporting its reuse by individuals throughout their life cycle. The capacity of computational systems to find, access, interoperate, and reuse data, with none or minimal human intervention, is essential in today’s data-driven era, where humans increasingly rely on computational support to deal with data as a result of the increase in volume, velocity and variety and complexity.
Here is a simple overview of what the FAIR principles recommend. In breif, data should be:
Findable: the first step in (re)using data is to find them, and descriptive metadata is essential.
- F1. (Meta)data are assigned a globally unique and persistent identifier
- F2. Data are described with rich metadata (defined by R1 below)
- F3. Metadata clearly and explicitly include the identifier of the data they describe
- F4. (Meta)data are registered or indexed in a searchable resource
Accessible: to get data one needs to know if authentication and authorisation is necessary, or if data is open with no restrictions.
- A1. (Meta)data are retrievable by their identifier using a standardised communications protocol
- A1.1 The protocol is open, free, and universally implementable
- A1.2 The protocol allows for an authentication and authorisation procedure, where necessary
- A2. Metadata are accessible, even when the data are no longer available
Interoperable: data need to be integrated with other data and/or interoperate with applications or workflows.
- I1. (Meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation.
- I2. (Meta)data use vocabularies that follow FAIR principles
- I3. (Meta)data include qualified references to other (meta)data
Reusable: data should be well-described so that they can be used or replicated in different settings.
- R1. Meta(data) are richly described with a plurality of accurate and relevant attributes
- R1.1. (Meta)data are released with a clear and accessible data usage license
- R1.2. (Meta)data are associated with detailed provenance
- R1.3. (Meta)data meet domain-relevant community standards
It is important to stress that making data ‘FAIR’ is not the same as making it ‘open’ (as accessibility principle clearly explains). Data should be as open as possible, as closed as necessary.
The FAIR principles refer to three types of entities: data (as any digital object), metadata (information about that digital object), and infrastructure (such as software and repositories). For instance, the findability principle F4 defines that both metadata and data are registered or indexed in a searchable resource (such as a data repository). For example, the FAIR principles are also being applied to software, and projects where the data and software are both FAIR the research is more likely to be reproducible.
It is much easier to make data FAIR if you plan to do this from the beginning of your research project. One way to do this is to create a Data Management Plan (DMP), in DMPonline or just as a text file, to help you think through how to manage your data. The DMP should include information on data creation (volume, formats/types and workflows), data use (where the raw or ‘live’ data is being stored), data publication and data archiving at the end of the project (long-term data storage, or what data is ‘kept’ at the end of a project). DMPs should also regularly be updated as the research project progresses or diverge from the initial design.
Tools and indicators
Altought started by a community operating in the life science, the FAIR principles have rapidly been adopted by publishers, funders, and pan-disciplinary infrastructure programmes and societies, in all disciplines. Many groups and organization are working to define guidances and tools to help researchers, as well as other stakeholders (such as librarians, funders, publishers, and trainers) to make data FAIR and assess its FAIRness level.
This rapid uptake and community involvement, however, has also caused some confusion and ambiguity on what FAIRness is and how we can measure it. It is important to say that the FAIR principles are aspirational, in that they do not strictly define how to achieve a state of FAIRness, but rather describe a continuum of features, attributes, and behaviors that will move a digital resource closer to that goal.
Listing all efforts working in and around FAIRness is practically impossible, as this is a fast moving, disperse and diverse field. Nevethless, if you are interested in following the discussion and/or participate in it, here are two global and international initiatives that act as umbrella organizations and reference points for many discipline specific efforts: GOFAIR and the Research Data Alliance (RDA). Under GOFAIR there are many Implementation Networks (INs) committed to implementing elements of the Internet of FAIR Data and Services within the three pillars: GO Build (Technology), GO Change (Culture) and GO Train (Training). Under the RDA there are several groups tackling different aspects relevant to the RDM life cycle, and among these one group is reviewing existing efforts, building on them to define a common set of common assessment criteria for the evaluation of FAIRness. Watch this space!
Metadata and identifiers - community standards
Metadata (to describe and report the data) and unique persistent identifiers (to cite and reference data) are the two pillars of the FAIR principles. The use of community-defined standards for metadata and identified is vital for high-quality, reproducible research and for the integrative analysis and comparison of heterogeneous data from multiple sources, domains and disciplines. Although also in this areas there is a lot of work in progress, and expecially metadata standards are disciplines specific, or specific to a given digital object, you need to know what these are and which one are relevant to your data type in order to mention them in your DMP and use them.
You can use FAIRsharing as a lookup resource to identify and cite the metadata and/or identifier schemas, databases or repositories that exist for your data and discipline, for example, when creating a data management plan for a grant proposal or funded project; or when submitting a manuscript to a journal, to identify the recommended databases and repositories, as well as the standards they implement to ensure all relevant information about the data is collected at the source. FAIRsharing also operates under GOFAIR and the RDA, and is widely adopted by publishers, funders, and other organizations; like any other FAIR-enabling resources, it will continue to evolve, better linking to DMP and FAIRness assessment tools, to better help you maing data FAIR a reality.
Storage and backup
Data loss can be catastrophic for your research project and can happen often. You can prevent data loss by picking suitable storage solutions and backing your data up frequently.
Where to store data
- Most institutions will provide a network drive that you can use to store data.
- Portable storage media such as memory sticks (USB sticks) are more risky and vulnerable to loss and damage.
Especially if you are handling personal or sensitive data, you need to ensure the cloud option is compliant with any data protection rules the data is bound by. To add an additional layer of security, you should encrypt devices and/or files where needed.
Your institution might provide local storage solutions and policies or guidelines restricting what you can use. Thus, we recommend you familiarise yourself with your local policies anc recommendations.
When you are ready to release the data to the wider community, you can also search for the appropriate databases and repositories in FAIRsharing, according to your data type, and type of access to the data. Major publishers are progressively defining clearer recommendations for data deposition and sharing, endorsing specific generics as well as domain-sepecific databases and repositories.
To avoid loosing your data, you should follow good backup practices.
- You should have 2 or 3 copies of your files, stored on
- at least 2 different storage media,
- in different locations.
The more important the data and the more often the datasets change, the more frequently you should back them up. If your files take up a large amount of space and backing up all of them would be difficult or expensive, you may want to create a set of criteria for when you back up the data. This can be part of your data management plan.
Data organisation in spreadsheets
Spreadsheets, such as Microsoft Excel files, are commonly used to collect, store, manipulate, analyse, and share research data. Spreadsheets are convenient and easy-to-use tools for these tasks but are not amenable to reproducibility if used as dynamic documents. There is a collection of horror-stories that document the many ways that the use of spreadsheets have scuppered analyses due to unexpected behaviour or error-prone editing processes. Some of these mishaps are not unique to spreadsheets, but many are.
Data manipulation and analysis in spreadsheets in particular is best avoided as, without version control, it can lead to non-reproducible workflows. By opening and editing raw data files directly by hand, for example to change values or perform calculations, the process by which new values are obtained is not properly documented, and you may accidentally over-write something or type in the wrong value only to notice after it’s too late (or not at all).
It’s often not practical to avoid the use of spreadsheets altogether but there are some simple steps that can be taken to mitigate their flaws. The following principles, taken from Broman et al. (2018), provide some practical advice to ensure your data is clearly organised and human- and machine-readable:
- Be consistent
- Write dates as YYYY-MM-DD
- Don’t leave any cells empty
- Put just one thing in a cell
- Organize the data as a single rectangle
- Create a data dictionary
- Don’t include calculations in the raw data files
- Don’t use font color or highlighting as data
- Choose good names for things
- Make backups
- Use data validation to avoid data entry mistakes
- Save the data in plain text files
Documentation and metadata
Having data available is of no use if it cannot be understood. For example a table of numbers is useless if there are no headings which describe what the columns/rows contain. Therefore you should ensure that open datasets include consistent core metadata, and that the data is fully described. This requires that all documentation accompanying data is written in clear, plain language, and that data users have sufficient information to understand the source, strengths, weaknesses, and analytical limitations of the data so that they can make informed decisions when using it.
- The level of documentation and metadata will vary according to the project and the range of people the data needs to be understood by
- It is best practice to use recognised community metadata standards to make it easier for datasets to be combined. For example, for brain data the Brain Imaging Data Structure is the strandards to use; other metadata standards(reporting requirements, termonologies, and models/schemas) are searchable in FAIRsharing.
- Variables should be defined and explained using data dictionaries
- Data should be stored in logical and heirarchical folder structures with a README file used to describe the structure. The README file is helpful for others and will also help you find your data in the future (Fuchs & Kuusniemi, 2018).
Sharing and archiving data
Motivations for sharing data
The world is witnessing a significant global transformation, facilitated by technology and digital media, and fueled by data and information. This transformation has enormous potential to foster more transparent, accountable, efficient, responsive, and effective research. Only a very small proportion of the original data is published in conventional scientific journals. Existing policies on data archiving notwithstanding, in today’s practice data are primarily stored in private files, not in secure institutional repositories, and effectively are lost to the public. This lack of access to scientific data is an obstacle to international research for two main reasons:
- It is generally difficult or impossible to fully reproduce a scientific study without the original data.
- It often causes unnecessary duplication of research efforts; large amounts of research funds are spent every year to recreate already existing data. Furthermore, it inhibits joint research activities on various aspects of the same problem.
Accordingly, there is an ongoing global data revolution that seeks to advance collaboration and the creation and expansion of effective, efficient research programs. Sharing data openly is crucial to meeting these objectives. Open data means that the data is freely available on the internet permitting any user to download, copy, analyse, re-process, and re-use it for any other purpose without financial, legal, or technical barriers other than those inseparable from gaining access to the internet itself. This represents a real shift in how research works. At the moment anyone who wishes to use scientific data from a researcher often has to contact that researcher and make a request. Open by default turns this on its head and says that there should be a presumption of publication for all. Researchers need to justify data that’s kept closed, for example for security or data protection reasons. Free access to, and subsequent use of, data is of significant value to society and the economy, and that data should, therefore, be open by default. Research is often publically-funded, so the results of this research should be openly available as a public good.
Steps to share your data
Step 1: Select what data you want to share
Not all data can be made openly available, due to ethical and commercial concerns, and you may decide that some of your intermediate data is too large to share. So you first need to decide which data you need to share for others to be able to reproduce your research.
Step 2: Choose a data repository or other sharing platform
Data should be shared in a formal, open, and indexed data repository where possible so that it will be accessible in the long-run. Suitable data repositories by subject, content type or location can be found at Re3data.org and in FAIRsharing where you can also see which standards (metadata and identifier) the repositories implement and which journal/publisher recommend them. If possible use a repository which assigns a DOI, a digital object identifier, to make it easier for others to cite your data.
Step 3: Upload your data and documentation
In line with the FAIR principles outlined above upload the data in open formats as much as possible and include sufficient documentation and metadata that someone else can understand your data.
Step 4: Choose a licence and link to your paper and code
So that others know what they can do with your data you need to apply a licence to your data. The most commonly used licences are Creative Commons, Open Government Licence, or an Open Data Commons Attribution License. To get maximum value from data sharing make sure that your paper and code both link to your data, and vice versa, to allow others to best understand your project.
Barriers to data sharing
Some academics still find sharing data difficult. Recent surveys (“Whitepaper: Practical challenges for researchers in data sharing,” 2018) amongst researchers find that sharing data is challenging for the following reasons:
- Organising data in a presentable and useful way (mentioned by 46%),
- Unsure about copyright and licensing as mentioned by 37%,
- Not knowing which repository to use as raised by 33%.
These are cultural challenges that might be addressed in changing practice going forward. However, there are also legal, ethical or contractual reasons that sometimes prevent making data publicly available in its entirety or even in parts. Those will be addressed in the following paragraphs.
Privacy and data protection
Many fields of scientific disciplines involve working with sensitive personal data. Their management is well regulated in data protection legislation (in Europe through national implementations of the General Data Protection Regulation) and ethics procedures as they are established in most research institutions (Regulation, 2016).
In order to make sure that anonymised research data can be made available for future reuse, it is important that consent forms cover sharing anonymised data with other researchers. Research so far suggests that study participants are usually less concerned about the data being archived and shared than researchers think (Kuula, 2010). Participant information sheets and consent forms should include how research data will be stored, preserved and used in the long-term, and how confidentiality will be protected when needed.
Individuals must be protected from (re)identification via their personal data used for research. Anonymisation of the data may be sufficient in some cases, but ensuring that reidentification is not possible is becoming increasingly difficult and might be impossible due to technical progress, growing computational power and – ironically – more open data. For example reidentification may be possible via data-mining of accessible data and so-called quasi-identifiers, a set of (common) properties that are – in their combination – so specific that they can be used to identify. Preserving privacy may still be possible if partial or generalised datasets are provided. For example, providing age bands instead of birth date or only the first two digits of postal codes. It may also be possible to provide the data in such a format that researchers can query it whist keeping the data itself closed. For example, a user may be able to send a query to retrieve the mean of a dataset, but not be able to access any of the individual datapoints. Another way to provide anonymized data is to provide synthetic data, data generated to reflect the conditions and properties of the raw data without including any of the personal information.
National and commercially sensitive data
In many cases companies are understandably unwilling to publish much of their data. The reasoning goes that if commercially sensitive information of a company is disclosed, it will damage the company’s commercial interests and undermine competitiveness. This is based on the thinking that in competitive markets, innovation will only occur with some protection of information: if a company spends time and money developing something new, the details of which are then made public, then its competitors can easily copy it without having to invest the same resources. The result is that no-one would innovate in the first place. Similarly governments are often unwilling to publish data that relates to issues such as national security due to public safety concerns. In such cases it may not be possible to make data open, or it may only be only possible to share partial/obscured datasets as outlined in the section above on privacy.
Susanna-Assunta Sansone: from FAIR co-author to FAIR doer
Let’s start with my digital footprints so that you can discovery my activities:
- The Data Readiness group I run at Oxford University
- ORCID: 0000-0001-5306-5690
- Profile on LinkedIn
- Talks on SlideShare
- Twitter; yes I am the FAIRlady
- Activities on GitHub; yeah, no code sorry
- Google Scholar
I have worked on research data management since 2001, yes, way before this area was even considered a ‘thing’. I was also told that there is no career to make in research data management. Well, how wrong that comment was! Formely seen as intersection between service provision and IT, data management has become a first class citizen, recognized as a research and development subject, as it should be. A lot of the credit for this change goes to the FAIR principles. Love it or hate it, FAIR is now an internationally-known lighthouse brand. Since we published the first article on the FAIR principles, enabling FAIR has been my focus.
The reuse of other people’s data is providing useful insights for new research questions and products, and driving new scientific discoveries. To realize its potential, however, we need new mechanisms to manage the growing availability and scale of scholarly digital products, such as datasets, software, algorithms, articles. FAIR has been specifically designed to emphasise the machine-readablity of these digital objects.
With my group of research software and knowledge engineerings we address the grand challenges related to information science and scholarly communications, where data quality and readiness for (re)use is a prerequisite for success. I believe that better data means better science, and this underpins reusable research, aids scholarly publishing, and enables faster and reliable data-driven discoveries. As I say in my one minute video, my vision is to transform the concept of data readiness into a powerful toolkit at the researchers’ fingertips, to realize FAIR data by stealth.
FAIR is not a magic wand. There is a lot to be done to enable and enact this transformation. We need all hands on deck! Researchers, service providers, journal publishers, library science experts, funders and learned societies in the academic as well as in the commercial and governmental setting all play a role: from providing use cases, to drive policy and culture changes that motivates, rewards and credits researchers for disseminating and publishing high-quality, machine-readable data; to building tools and services, to inform, training and educate.
There are many community efforts around FAIR; keeping abreast with these is an activity in itself. I spend considerable time to bring my group activities (such as ISA and FAIRsharing) in and under larger international umbrella organizations like GOFAIR and the RDA to interact with others, learn from them, compare and contrast efforts and build new collaborations. I also play leading roles, seating on boards and chairing working groups with collagues, because you must get your hands dirty and lead by example.
In research data management, the history is the future. The one I envision is a future where scientific evidences are routinely available in a transparent, trustworthy and persistent manner to support peer-review and withstand reproducibility, to underpin new results and discoveries, and effectively drive sciences forward.
My advice: be aware, be FAIR!
Don’t Touch the Raw Data - back it up somewhere reasonable and keep a read-only copy.
Have a plan! - Decide where your data is going to be stored, what its called, when/if it needs to be deleted BEFORE you start collecting it and note it down in a data management plan. If you collect sensitive data, plan for consent for sharing from the start!
Document Everything - You know who the worst person at replying to emails is about what the sampling frequency of channel X was. Nope not him, it’s actually yourself from a year ago. Keep the documentation with the data!
Create the data you want to see in the word - Imagine someone was about to give you a dataset that you needed to analyse well in order to get that job you’ve been dreaming about. What would you want it to look like? That’s how yours should look.
Try not to re-invent the wheel. Before you start creating some crazy new schema, storage format or naming protocol - have a quick google, ask your colleagues. Something that’s already being used is likely going to be better in the long run even if you think there’s a better solution.
What to learn next
- Detailed guidance on sharng personal or sensitive data from the UK Data Service
- An overview of storage solutions and their advantages and disadvantages
- Wilkinson, M. D., Dumontier, M., Aalbersberg, I. J. J., Appleton, G., Axton, M., Baak, A., … Mons, B. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. Retrieved from https://doi.org/10.1038/sdata.2016.18
- Broman, K. W., & Woo, K. H. (2018). Data organization in spreadsheets. The American Statistician, 72(1), 2–10. https://doi.org/10.1080/00031305.2017.1375989
- Fuchs, S., & Kuusniemi, M. E. (2018). Making a research project understandable - Guide for data documentation (Version 1.2). Zenodo. https://doi.org/10.5281/zenodo.1914401
- Whitepaper: Practical challenges for researchers in data sharing. (2018). Figshare. https://doi.org/10.6084/m9.figshare.5975011.v1
- Regulation, P. (2016). Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation). Official Journal of the European Union. Retrieved from http://data.europa.eu/eli/reg/2016/679/oj
- Kuula, A. (2010). Methodological and Ethical Dilemmas of Archiving Qualitative Data. IASSIST Quarterly, 34(3-4), 35. Retrieved from http://www.iassistdata.org/sites/default/files/iqvol34_35_kuula.pdf