Overview
Teaching: 15 min Exercises: 5 minQuestions
How can I make my own data available to others?
How should I cite datasets?
Objectives
Explain the publication process
Explain DOIs, ORCID, Zenodo
Data management plans
Metadata is important for others to be able to understand and use your data. The standard in the climate area is CF-NetCDF, which has standard variable and dimension names that make it possible to compare results from different centres agains each other
Without metadata it is much harder to get tools like THREDDS to work, harder for other researchers to use - what are the units for this field?
It’s easier to collect information as you go, tools will generally preserve attributes. Saves having to backtrack through everything at the end
Metadata is for both single files as well as groups of files
If the right data is present (even if unstructured) data can be published in a matter of days
Attribute convention for data discovery is a good list of recommended attributes
History give a view on how the file was created, use timestamps
NCO’s ncatted can edit attributes
ncatted --attribute NAME,VARIABLE,MODE,TYPE,VALUE input.nc output.nc
Example: CF-netCDF metadata
To publish our NINO 3.4 data we’ll first need to add some metadata
When you apply for a grant from the ARC you may need to provide a data management plan, which provides information on where your data is stored and how it’s made available to other researchers
Institutions also require data management plans, especially for students - thesis related data needs to be kept in an institutional archive for a number of years after graduation.
Think about licencing, creative commons is ANDS recommendation
How do you manage your own data?
Think about your own project * Will data need to be published when you publish papers or apply for grants? * Where is your data stored? * Is it backed up? * How long can it stay there? * Do you have documentation and metadata alongside your files? * Is your data licensed?
Papers, datsets and model code can be assigned a permanent identifier, called a DOI (digital object identifier) so that others can cite it.
You can follow a DOI link to the canonical location using the website https://www.doi.org/
Papers are assigned DOIs by the publisher. NCI can arrange for DOIs for data published and served on their systems, your institution may be able to help as well
ANDS partners with ORCID, a researcher ID system, to match datasets with authors
ARC encourages researchers to sign up
Not-for-profit organisation linking researchers with publications and data
Code is also data, and important to preserve and cite! Keep your code in version control (github, bitbucket, gitlab). You can assign a DOI to a specific code version, either through ANDS or tools like CERN’s Zenodo
Recommended citation formats
Creator (PublicationYear). Title. Publisher. Identifier
Creator (PublicationYear). Title. Version. Publisher. ResourceType. Identifier
ARCCSS CMS are happy to help publish your data. We have a dedicated data manager who works with NCI and ANDS to make sure your data is available and citable, email climate_help@nci.org.au
Key Points
Publishing datasets can be a requirement of grants or journals
CMS is happy to assist with publication - climate_help@nci.org.au