For the second case, researchers and developers are working hard to build options for capturing the computing environment in which data is built or analyzed. For help in finding those tools, take a look at resources like ReproMatch, a searchable list of reproducibility tools. Or give tools like ReproZip a try. ReproZip is a tool developed by the ViDA (Visualization and Data Analysis) group at NYU to make reproducibility easy. It tracks operating system calls and creates a package that contains all the binaries, files, and dependencies required to run a given command on the original researcher’s computational environment. A reviewer can then extract the experiment in his own environment to reproduce the results, even if the environment has a different operating system from the original one.

Another good place to start to follow developments in using and re-using data is Reproducible Science, a consortium arising out of the Moore/Sloan Data Science Environment and its partners, New York University, the University of California at Berkeley, and the University of Washington.

Being able to cite data, and to track how and where that data is getting cited, is also becoming a visible aspect of publishing. When seeking a permanent home for your dataset in conjunction with results publication, take great care to seek a repository that will make your data available in a way that is advantageous to citation and discovery.

Make sure you understand how your data will be licensed for use, and verify that the repository has all of the proper elements such as DOI provisions, download and even citation tracking, and a good balance between visibility within your discipline and discoverability outside of your scholarly circle. In other words, all of the issues a researcher would consider when seeking a place to publish their written reports, articles, and analyses!

——
¹ Yvonne M. Sorcha, ed., “Out of Cite, Out of Mind: The Current State of Practice, Policy, and Technology for the Citation of Data,” Data Science Journal 12 (13 Sept. 2013): 3.2.1 <http://dx.doi.org/10.2481/dsj.OSOM13-043>, and M. Martone, ed., Data Citation Synthesis Group: Joint Declaration of Data Citation Principles (San Diego, CA: FORCE11, 2014) <https://www.force11.org/group/joint-declaration-data-citation-principles-final>.

Undoubtedly, the answer is largely related to time, a scarce commodity in the current pressured production environment for research. And one cannot avoid the fact that it does take a certain amount of time and effort to get a documentation system off the ground. The good news, however, is that documentation systems often demand close attention and resources early in the process, but once a workable system is in place it can be replicated across projects in a way that makes the effort less burdensome down the road. And good research management is often about increasing efficiency. Use automated computing processes whenever possible to reduce the time in manipulating research files. Deploy central dashboards that bring together the various threads of your work. And if you work on multiple devices in multiple locations (as we often do), get your cross-device working environment in order so that you don’t have to constantly reformat work done on one device so that it conforms with the official documentation scheme you have deployed on another.

Some common practices deployed in documenting data include:

• Use of readme files to explain the contents of folders and directories. Take a look at this helpful guide from Cornell University’s Research Data Management Service Group for ideas.
• Reliance on a project management dashboard, e-Lab Notebook, or cloud-based file manager to gather everything together. We list several on our Documenting Data guide.
• Matching of long-term documentation systems to ongoing research analysis needs as far as possible. In other words, use the same strategies for documenting your research materials in the long run that you will also use to push your analysis forward and enable publication.

Happy documenting, and keep up with Wednesday’s Love Your Data week focus on documenting data on Twitter, #LYD16!

Documentation, systematic file naming, and good computer file management are all essential to avoid this problem. Follow along at #LYD16 on Twitter for reminders about best practices in selecting file names and recommendations for software to document files. You can also check out the Research Data Management library guide section on file names at guides.nyu.edu/data_management/file_org.

To avoid this, you need to have your data in two different physical locations and on two different media. This is accomplished by having your data saved in three places: on two hard drives and a cloud, or on a hard drive and two cloud instances.

But remember that a good storage environment needs updating. You data is probably constantly being changed and versioned. So a good storage setup also includes robust provisions for tracking versioning and storing older instances of files. Consider setting up an automated backup schedule, a common feature on many external hard drives, and look for cloud services such as NYU Box that track file changes. Most importantly, work out a file version naming system and document the contents of data files as you go.

For more information about file storage options and practices at NYU, consult the Research Data Management library guide.

Monday: We will be looking at ways to keep your data safe through tips on storage, formats, and navigating a cloud environment.

Tuesday: Keeping your data organized and accessible!

Wednesday: What are the best approaches to documenting your data? Check in with #LYD16 on Wednesday for more options.

Thursday: With research verification and reproducibility increasingly featured in today’s scholarly publishing, ensure that you can give and get credit for your data.

Friday: Data resusability. Learn how your data can be transformed so that it is usable by you and by other researchers in the future.