Big Changes in Data Ahead

The White House made an announcement on Friday that will significantly impact the way that you disseminate your research. The memorandum (pdf), covering all federal granting agencies with over $100 million in annual R&D expenditures (a list of agencies can be found in this Scholarly Kitchen blog post), directs granting agencies to make the products of their grants–both publications and data–accessible to the general public. Here’s an overview of the announcement and what it means for you, the researcher.

Changes in Publications

The publication portion of this directive is akin to what the NIH currently mandates. Articles will be published through normal channels but made freely available to the public after a 12-month embargo. The NIH houses these publications in PubMed Central, though I expect that other granting agencies will make use of different repositories. What this means is that in the short term you will publish articles as normal but in the long term more people will be able to read your articles.

Changes in Data

The publications announcement alone is really big news but, as this is a blog about data, I’m more interested in the data portion of the memo. Here’s where things get a little more complex because there isn’t nearly as much infrastructure to support data sharing.

The memo does not give an outright mandate for data sharing, instead it seeks to “maximize access” to digital research data. I definitely view this as a step in the direction of full sharing; it gives time for researchers to adjust to the new model and for the rest of us to build infrastructure and work out the kinks.

It should be noted that even this ‘maximized access’ has it’s limitation. Among other things, researchers will not be forced to share classified data, data with privacy and confidentiality concerns, lab notebooks, preliminary analysis, and physical objects (see memo section 4 for the full list). Researchers can also exert intellectual property rights over data to prevent sharing and I’m looking forward to seeing how this manifests as policy.

Beyond encouraging data sharing, the memo’s directives model the current NSF mandate for data management plans. All researchers on federal grants will be required to create data management plans and agencies should monetarily support these plans, as appropriate.

How the memo effects you as a researcher is less clear for data than for publications. It is clear that you will be writing more data management plans and more resources will likely be put into supporting and evaluating these plans; I expect research institutions will be stepping up in this area as well as funding agencies. Researchers will also start feeling more pressure to share their digital data, but avenues for sharing data will increase and sharing (and citing) datasets will become easier over time. Any further requirements on researchers will become defined once the agencies develop their data policies.

Time Scale

Granting agencies will have 6 months from the time of publication of the White House memo to develop new policies that comply with the memo. These agencies will also need to do this work within their existing budgets. For these reasons, the changes outlined in the memorandum will not happen right away but you should be aware that they are coming soon.

My Final Thoughts

The White House memo is a measured step toward opening up research because it makes use of existing NSF and NIH policies that have been proven feasible. For this reason, I expect the publication and data management plan portions of the memo to proceed somewhat smoothly. As for data sharing, we are at the beginning of a big change in what we do with digital research data and it will take some time to settle into a new system of sharing. I don’t think that this memo is the last word in data sharing and any further changes are likely to be as measured and deliberate as this memo.

In the end, I am personally very excited about the White House announcement and will spend a future blog post discussing why opening up research is a good thing for science.

Posted in fundingAgencies, government, openAccess, openData | 2 Comments

Starting Small: File Naming Conventions

Managing research data well can feel like an overwhelming task, so it’s important to start small. It’s much easier to make several small changes over time than to change the whole system at once.

One of the easier and more helpful changes you can make in the laboratory is to utilize consistent file naming. A file naming convention add standardization to your files, making them easier to organize and locate. It will also help your coworkers sort through your files should you fall ill or leave your job; your naming scheme should be documented in your laboratory notebook (preferably at the front or back for easy access) for this reason.

There are a lot of conventions available for you to choose from, though you will probably want to customize one for your own purposes. Here are a few general tips for naming files:

The Basics

  • Files should be named consistently
  • Files names should be descriptive but short (<25 characters)
  • Use underscores instead of spaces
  • Avoid these characters: “ / \ : * ? ‘ < > [ ] & $

Dates

  • Date your files using the convention YYYY-MM-DD

I think that dating your files is one of the best ways to help organize your data, particularly because paper lab notebooks are organized by date. So if you have only the file, you can look at the date in its name and immediately know where to search for the corresponding notes in your notebook. (You can also reinforce this file-notebook connection by organizing your computer’s folders with reference to each notebook.)

Versions

  • For analyzed data, use version numbers
  • Save files often to a new version
  • Label the final version FINAL

Versioning can be imminently helpful when you are manipulating data. If you make a change to your data that you don’t want to keep, it’s simple to go back to an earlier version of the file. The same is true if a file gets corrupted or if you simply want to change your analysis method. The key to making versioning work is being consistent with version names, periodically saving to new versions, and documenting the differences between versions.

These are some general thoughts on file naming. Feel free to leave a comment about a system that has worked well for you!

 

Sources:

http://researchdata.wisc.edu/manage-your-data/file-naming-and-versioning/

Posted in digitalFiles | Leave a comment

The Blog I Wish I Had

When I was a practicing chemist, I wore many hats. One day I was a programmer, the next a plumber, and only sometimes was I actually a chemist. While I enjoyed the day-to-day variety, I found it hard to adapt to roles in which I had very little training. One such area was dealing with my research data.

As an undergraduate, I learned a little bit about keeping a laboratory notebook, but this wasn’t enough to prepare me for managing and organizing my data once I got to graduate school. Left on my own, I cobbled together a system that wasn’t very good and was absolutely no help to anyone else. I knew my system could be better but I didn’t have the knowledge to actually make it better.

I’ve learned a lot about dealing with data since then and I wish I could go back and share that information with my former self. Instead, I’m sharing that information with you.

My plan for this blog is to write about managing research data in the lab, in addition to discussing how data are becoming first class citizens in the world of research. There are a lot of exciting things going on around research data at the moment and I can’t wait to talk about them with you!

Posted in admin | Leave a comment