10 Tips for better data for Love Data Week
1. Get your protocols organized
Some are on Google Drive, some are photocopies from a previous lab member's handwritten notebook, and some were jotted down on a napkin over coffee and have never been written out officially. Get your methods, protocols, and computational workflows organized using protocols.io. Stanford's premium license allows you free access to this web-based service where you can create an unlimited number of private protocols, then share them with your research group and refer to them directly in publications via their very own DOIs (digital object identifiers). Find out more.
2. Implement a file naming system
There's no one way to set up a file naming system -- it depends on the work that you do. But most of us could certainly do better. Check out our best practices for file naming (and a great example included there) to learn the components of a good system, and then set yourself up with a plan. You might even want to rename your existing files so that you'll be motivated to use and maintain the new system in the future (just make sure you update any references you have to those files).
3. Re-evaluate the file formats you use
If you are going to want to share your awesome research in the future (and we hope you will), then you'll want to do it in the most accessible format. Sometimes it's difficult or even impossible to convert files from a proprietary format into something that is open, but tables in Excel -- if done well and with an eye toward future sharing -- can easily be converted to .csv or similar accessible formats. We have some thoughts on best practices for file formats to share with you (you knew we did, didn't you?).
4. Implement a versioning practice
I know, I know. You keep meaning to do this. It just seems like such a hassle. But we have some tips on data versioning that can get you started with baby steps. I spend a fair amount of time trying to make sure that the future me doesn't regret having not done something. Trust me, the future you will wish you'd learned to use version control.
5. Metadata, metadata, metadata
Learn what metadata you should be collecting about your research data and then set up a method for doing this. We have info on creating metadata -- from basic methods to more complex -- as well as some tools that may fit your needs and preferences. A readme.txt file may be a good place to start.
6. Figure out what you'll share
Eventually, you'll either want to or be required to share your research data. It makes sense to start sorting this out as early as possible. That whole set of runs where you found out later the equipment wasn't properly calibrated? Not something you're likely to need (but you never throw anything out, am I right?). But you'll want to be able to find all the important pieces. Finding it now and setting up good ways for tracking this in the future is an excellent use of time. See our tips on selecting data for sharing and preservation, as well as specific guidance for sensitive information.
7. Learn about licenses
When you share your research data, it's best practice to assign it a license. That way people will know exactly what you are allowing them to do with your stuff. No license? Then people have no idea and may assume they can do whatever they want. Learn more about licenses that are frequently attached to research data and start thinking about which one is best for you and your data.
8. Deposit your data in the Stanford Digital Repository
If you have data that you are ready to share now -- in conjunction with a grant or a publication or because it's something useful for others that you just want to share -- request access to the Stanford Digital Repository (SDR), a service offered by Stanford Libraries. Most data can be uploaded in 15 minutes or less via our online deposit application. Contact us for access to the SDR.
9. Get a DOI
If your publisher or funder requires a digital object identifier, or DOI, for your research data, then you can get one with the click of a button when you deposit data in the SDR (see above). We even have an option to request this first -- before you actually deposit anything -- so you can include in your manuscript the link for where the data *will* be in the future. If you need a DOI but your data are not appropriate for the SDR, we may still be able to make one for you, so contact us.
10. Ask us
Questions? We're here to help! Contact us at ask-data-services@lists.stanford.edu.
