Data Management for
Office of Undergraduate Research Seminar and Workshop Series
Rebekah Cummings, Research Data Management Librarian
J. Willard Marriott Library, University of Utah
June 18, 2015
• What are data?
• Why manage data?
• Data Management Plans
• File Naming
• Storage and Archiving
What are data?
“The recorded factual material
commonly accepted in the research
community as necessary to validate
- U.S. OMB Circular A-110
Data are diverse
Data are messy
Why manage data?
Your best collaborator is yourself
six months from now, and your past
self doesn’t answer emails.
Why else manage data?
• Save time and efficiency
• Meet grant requirements
• Promote reproducible research
• Enable new discoveries from your data
• Make the results of publicly funded research
We are trying to avoid
Two bears data
1. Didn’t know where he stored the data
2. Saved one copy of the data on a USB drive
3. Data was in a format that could only be read by
outdated, proprietary software
4. No codebook to explain the variable names
5. Variable names were not descriptive
6. No contact information for the co-author Sam Lee
Data Management Plan
Courtesy of the UK Data
You develop a research project during your
undergraduate experience.You write up the
results, which are accepted by a reputable
journal. People start citing your work! Three
years later someone accuses you of falsifying
Scenario adapted from MANTRA training
• Would you be able to prove you did the
work as you described in the article?
• What would you need to prove you hadn’t
falsified the data?
• What should you have done throughout
your research study to be able to prove
you did the work as described?
Elements of a DMP
• Types of data, including file formats
• Data description
• Data storage
• Data sharing, including confidentiality or
• Data archiving and responsibility
• Data management costs
File naming best
• Be descriptive
• Don’t be generic
• Appropriate length
• Be consistent
Who filed better?
File naming best practices
• Files should include only letters, numbers, and
• No special characters (%@#*?!)
• No spaces
• Lowercase or camel case (LikeThis)
• Not all systems are case sensitive.Assume this,
THIS, and tHiS are the same.
Dates and numbering…
1. Use leading zeros for scalability
2. If using dates use YYYYMMDD
June2015 = BAD!
06-18-2015 = BAD!
20150618 = GREAT!
2015-06-18 = This is fine too
Who filed better?
• July 24 2014_SoilSamples%_v6
File organization best
• Top level folder should include project title
• Sub-structure should have a clear and
consistent naming convention.
• Document your structure in a README
File organization exercise
There was a study put out by Dr. Gary
Bradshaw from the University of
Nebraska Medical Center in 1982
called “ Growth of Rodent Kidney
Cells in Serum Media and the Effect of
Viral Transformation On Growth”. It
concerns the cytology of kidney cells.
Title Growth of rodent
kidney cells in serum
media and the effect of
viral transformations on
Author Gary Bradshaw
Publisher University of Nebraska
Subject Kidney -- Cytology
Why create metadata?
• Interview protocols
• Lab notebooks
• Code or scripts
• Consent forms
• Samples, weights, methods
• Read me files
LOCKSS (Lots of
Options for data
• Personal computers or laptops
• Networked drives
• External storage devices
Storing sensitive data
• If possible, collect the necessary data
without using direct identifiers
• Otherwise, de-identify your data upon
collection or immediately afterwards
• Do not store or share sensitive data on
• Talk to IRB
• Public repository – FigShare
• Domain-specific repository
• Institutional repository
• Data management starts at the beginning of
• Document your data so that someone else
could understand it
• Have more than one copy of your data
• Consider archiving options when you are
done with your project
Marriott Library, 1705Y
…or ask now!
Specifically we are going to be be talking about data management of your research data, but some of the principles will help you when thinking about the organization of any digital materials, your notes, your PowerPoints, your grocery lists….
. Most of these concepts are pretty straightforward, they almost seem like common sense, but the reality is that very few people manage their data well and if you do, you will be at a big advantage.
Overview of what we will be covering in this session. Each of these could be a one hour course, but we are going to hit the highlights so to speak.
Are you working on a research project?
What is data?
This is the definition that most people refer to.
Recorded factual material
Validate your research findings – when you write up your research it usually ends with your findings. What you discovered in the course of your research. Data is how you got there. It’s your proof.
Data are a lot more complicated than that OMB definition. Data is whatever you consider to evidence for the research that you do. In that way, data can be very subjective.
Scientific data – observations, computational models, lab notebooks
Social sciences – results of surveys, video recordings, field notes
Humanities – text mining, newspapers, records of human history
So what is data – EVIDENCE FOR YOUR RESEARCH
Another attribute of data is that it tends to get messy
Most of us just don’t realize this because our messy, disorganized files are locked up in a neat little box called your computer.
Don’t believe me? How long would it take you to find a photo from five years ago on your computer? Here is a hint. If your image files start with DSC_ or IMG_ and some number following it, it will probably take you a very long time.
If most people’s digital files were analog, this is exactly what they would look like.
The main reason you should manage your data is for yourself and for your own research team.
Data management is one of those essential skills you need to get just like learning how manage citations or understand research methods.
But it can feel a bit boring like filing. But six months later when you want to locate a file, or even understand your file, your future self will thank you.
Most important reason to have good data management is for your own good and the good of your research team. If you want to be able to locate your files or understand your files in the future, good data management is crucial. Plus, unlike research methods and managing citations, this is something that even seasoned scientists are not very good at. So you will have something to offer your research team in the future even as a young scientists.
For all the reasons we have talked about, many agencies are now requiring data management plans at the start of a research project. This means when you apply for funding for a project, you will have to have a two-page data management plan as part of your proposal. That plan is going to talk about the “lifecycle” of your data throughout the course of the project.
How many of you plan on applying for a grant at some point in your careers?
Introduce data lifecycle.
Funders know that the earlier you start thinking about your data, the better. It’s much more likely that the results of your research will be reproducible, it helps avoid data loss, and increases the value of your research.
Hopefully by now you can all see why data management is important. Now we’re going to think a little more deeply about how we can avoid the “Two bears” situation.
Let’s look at this scenario…
Get in groups and talk about this for a few minutes.
The first thing that you would want to have is a DMP. The DMP is going to be your roadmap for good data management. This is the document that you create at the start of a project to think about the lifecycle of your data.
We’ve talked about data management at kind of a high level. What is data? Why should you manage it well?
Now we are going to talk about some of the nuts and bolts of data management. Starting with file naming. How do you currently name files? Do you have a system?
To some extent we are all guilty of bad file naming but when it comes to your research it is important to create a system that makes sense not just to you, but other people as well.
are all guilty of bad file naming but when it comes to your research it is important to create a system that makes sense not just to you, but other people as well.
File names should reflect the contents of a file and enough information to uniquely identify the data file without getting way too long.
Don’t be generic in your file names
Your file name may include project acronym, location, investigator, date of data collection, data type, and version number. Whatever will help you or someone else uniquely identify that file in the future.
Think about what can be added and what can be omitted in your file names. If you are the only person on a project, you probably don’t need your name. If there are going to be multiple versions of a file, make sure you add a version number or a date to differentiate.
#1 is the best one.
Not too long, not too short
Nothing that makes it look like your file name is swearing at me.
Uppercase lettering can affect numbering.
There are also best practices around version control and numbering.
Version control is often achieved by using dates or a standard numbering system
#2 is the best choice here.
First example here has spaces, irregular dates that won’t line up in order, special characters
Third example may not be descriptive enough for for a secondary user. Also, beware of the “FINAL” as opposed to using a standardized numbering system.
That is how to name an individual file. What about your whole file structure?
All your research materials need to be in one folder. The top level folder should include the project title and year. If it is multiple year, include the first and last year in the title.
The substructures should have a clear and consistent naming convention that is documented in a README file.
Organize by type of file (all transcripts in one folder all audio recordings in another)
Organize by person (Have a Cliff Barrett folder and a Robert Bennett folder)
Problems with file names:
Dates are not standardized
File type in the file name which is unnecessary
Unnecessary information in file name – “found on Internet, think okay, better than mine” picture
NO consistency to file naming
Metadats is very, very important for other people looking to use your project.
Often called data about data.
Structured information about an object.
Mention that there are standards for creating metadata (Dublin Core) including subject specific data.
Data needs context to be understandable
If you have a spreadsheet of survey responses, you need to have the survey to understand the responses.
You also need the codebook that explains your variable names and the values that you used, how you cleaned your data. Once again, try to think how a secondary user would interpret your data.
Going back to file organization, make sure your data documentation is stored in the same folder as the data.
You must make a codebook and include it in your documentation.
This is documenting at a variable level. It’s just as important that you document at a Project and file level as well.
Summary, good data documentation includes…
Through the course of your research your data needs to be stored securely, backed up, and maintained regularly. Once again this sounds like common sense, but you will be happy when you pay some attention to it. (e.g. when your laptop crashes or is stolen.).
#1 rule of data storage – never just keep your data on one device. You are one dropped computer, one spilled glass of water, one unscrupulous thief away from losing all of your data. Every single day I go to Mom’s Café and see people leave their computers at their table while they go to the bathroom or grab a cup of coffee.
LOCKSS - There should never just be one copy of your data. Do you backup your data? Most important data management task. NO less than two, preferably three copies of research data.
How well are you covered against unexpected loss? Make sure that when disaster strikes, it isn’t a disaster
There are three options for
Personal computers and laptops – Convenient for storing your data while in use. Should not be used for storing master copies of your data.
Networked drives – Highly recommended. You can share data. Your data is stored in a single place and backed up regularly. Available to you from any place at any time. If using a department drive or Box stored securing thereby minimizing the risk of loss, theft, or authorized access. BEST!!!
External storage devices – thumb drives, flash drives, external hard drive. Cheap, easy to store and pass around. Feel better knowing it’s in your hands where you can see it. Not recommended for the long-term storage of your data.
Throughout the course of your research, many of you may collect data that is referred to as human subject data. If you do this, you will need to work with the IRB office on campus to figure out how to protect the privacy of your research subjects. Ultimately, the IRB has the final say, but here are some tips for keeping your confidential data, confidential.
Direct vs. Indirect identifiers
Another area of data management that you will have to consider is data archiving.
Archiving is not the same thing as storage
Archiving adds additional value to your data.
Sharable, usually through a persistent identifier
Makes data citable
There are lots of archiving options for your data. Some people choose to put their data on their website which is an option, but not a best practice.