Training Course on Data Management for Information Professionals and In-Depth Digitization Practicum September 2011, Oostende, Belgium Concepts of Data Management Planning and Mandates Lisa Raymond

So why do you need a DM plan? Beside the current trend of mandates from funding agencies, a good DM plan that makes data accessible will benefit the original researchers with more citations and it will benefit science and society by allowing for confident re-use of the data in decades to come.

In this era where data is driving science we have shifted from past practices of describing phenomena to collecting more data than we can possibly hope to archive over the long term. Good data management is essential to make informed decisions.

Elements of a Data Management Plan

What types of data will be created? Identify the material that will be created Model input Images Data samples Documents... Or perhaps you are building something – blueprints, software, calibrations, tools

What standards should be applied for format, metadata, etc. ? What are the natural ways the material to be created could be organized? By measurement kind Are there standards in your community – file type, etc. Scale does matter here

What standards should be applied for format, metadata, etc. ? Continued Topics to consider: Data gathering and QC Directory structures Identifiers and naming standards Versioning File formats Metadata standards Documentation

Where will the data be archived? Access, Sharing and Re-use Backups, archiving and preservation How are your ensuring your material are not lost Lots of Copies Keeps Stuff Safe (LOCKSS)

What provisions are there for access policies? Access, Sharing and Re-use What will be shared When and how Any restrictions

Are there plans for transition or termination of the data in the long-term future? How long do you need to keep Which material should be preserved for the long term Where is that going to happen When is that transfer going to happen

Metadata needed to understand data Metadata

Enabling data sharing Full life cycle Data first class citizen Value of outcome – not just outcome Data at heart of science informatics Rate of data collection will outstrip ability to keep it all

Role of Libraries Partner with data managers, researchers Action Items Informational Talks Survey Community Gather information Create tools to assist PI writing DM plans

Survey – see what PIs want/need Survey developed by a team: scientist, data manager, librarian and an informatics specialist. Designed as an education tool as well as gage of needs of community.

Data Management Survey Results Summary Sixty seven people responded to the survey, several indicated they were responding for their group. The Directorate/Program that is most often submitted to is Ocean Sciences, followed by Marine Geology & Geophysics and Physical Oceanography. The most common place to store data is on a local computer. Less than half of the respondents said they have a plan to manage data collected since More than half of the respondents said they would like assistance with a data management plan.

Survey Results One of the most important things to note is that the majority kept their data on a computer in their lab. This is not a good data management plan! The other important piece of information is that respondents did want help with data management plans --- an excellent opportunity for librarians to become involved.

Gather Information Survey Results Talk to other Librarians, view websites, join curation listservs, etc. Attend workshops like this one

Data Management Plan Checklist TALK TO YOUR PROGRAM MANAGER. The Directorate you are submitting to may have established protocols for an appropriate data management plan. Address solicitation requirements and any special requirements from the program. Your data management plan should identify the data products that will be created, the standards used and organization of the data-related materials. You should state the access, sharing and re-use policies and address backups, archives and preservation. We have created a checklist to help you describe the data plan, but no one list fits all research projects. Here are some points to consider as you write your data management plan Title - Data Management Plan for [grant title] Author(s)– PI, data manager, etc. Identify the data-related materials that will be created – for example experimental data, model output, remote sensing, software, etc. What are the natural ways the material to be created would be organized? What are the standards in your community? If there are community standards and you are not using them, why?

How will the data be gathered? Quality Assurance/Quality Control procedures? Directory structure Identifiers and naming standards Versioning Metadata standards Documentation Are there policies that define what data will be shared? What will be shared When will data be made available Where, is it publicly accessible Will it have a license – for example Creative Commons Any restriction on sharing or re-use, reasons?

How are you ensuring that your materials are not lost? How/when are you creating copies (Lots of Copies Keeps Stuff Safe) Where are the copies How long do you need to keep What data will be maintained beyond the life of the project? Which materials need to be preserved for the long term Where is that going to happen When is that transfer of data going to happen

Publicize Information

Mandates for Data Management Planning A lot of the current emphasis on data management plans is driven by mandates from funding agencies. Some had existing data management policies, but did not implement them. The increase in the capabilities of mining digital data and the huge amounts of data being Generated necessitate data management planning.

National Science Foundation (US) Data Management Policy Implementation The US National Science Foundation implanted their plan this year. There is little consensus on what exactly they want – the policy states that standards are dictated by the community. So Information Professionals are getting involved to help establish standards and best practices.

NERC Data Management Policy Implementation The United Kingdom based Natural Environmental Research Council does have a data policy handbook, the most recent dated The British Oceanographic Centre handles the Oceanographic data for NERC and is a partner in the pan European project – SeaDataNet, which has done a great deal of work on developing standards. They are a good resource for anyone working in the marine sciences.

Australian National Data Center Data Management Policy Implementation The Australian National Data Center provides guides for choosing metadata data standards and information on file formats, as well as persistent identifiers, copyright and citation.

Common Elements Data must not only be archived but be accessible. This ensures transparency of research and allows for the confident re-use of data in the future. Allows for data attribution and citation which gives credit to data originator.

Resources WHOI Data Management Site: ESIP Workshop: NSF FAQ: University of Oregon: Glossary of Metadata Standards mphlet.pdf mphlet.pdf CDL Guidelines for Digital Images SeaDataNet Australian National Data Center NERC