Data sharing and access
Enhancing data sharing and access through the R2K Data Portal
Moderator: Vicki Ferrini
This session will provide a brief update on new developments in the R2K Data Portal including the new search interface , and new functionality available in GeoMapApp. Please bring your laptop to follow along during some brief exercises intended to better familiarize you with Data Portal tools and functionality.
Data's shameful neglect (pdf)
http://www.nature.com/nature/journal/v461/n7261/full/461145a.html
Working lunch on including models in the R2K Data Portal
Moderator: Laurent Montesi
Background: As R2K moves more and more towards Integration and Synthesis, modeling (of all sorts) is become a more and more central activity of R2K projects. The NSF-R2K Data Management Policy requests that results be shared through the Data Portal, but what is there to do when models are the project results? This discussion is the opportunity for modelers to discuss what should and shouldn't be part of a data portal that is open to models and modeling results. Note that the discussion should look beyond R2K to the way that models are shared in Earth Sciences.
Participants (list your name here):
- Laurent Montési
- Roger Buck
- Arthur Olive
- Bill Lavelle
- Christol Miele
- Eun-seo Choi
- Garrett Ito
- Tim Crone
- Suzanne Carbotte
- Bob Lowell
- Craig Lundstrom
- Hélène Carton
- Meg Tivey
- Vicki Ferrini
Agenda:
What aspect of models should be in the data portal? Codes?
- No requirement to have the code itself, in part because it wil be impossible for some commercial software.
- Need contact info; Open up for NSF-funded software development, but leave developer choice of posting code or link to other site; Make clear the need to contact author to learn how to use the software or collaborate! Build up pages for each code, with a description, usage, and example applications. Input; description in code page/ need metadata for supporting data when relevant.
- Output: no general requirement. At least visualiztion; Can handle raw data; description in code page Separate start page to browse conceptual/general models that may not be georeferenced.
Model metadata
- Software name and version, contact info, key references
- Input file; link to input reference (database)
- Output description; guidelines for applications, sensitivity.
How can the available data and its organization be improved with modelers in mind?
Discussion topics (list ideas and suggestions here)
- One size does not fit all: What are the categories of models?
- What should be posted: full model outputs or example visualization? In what format?
- The source of the problem: do we post the software itself? Is metadata sufficient?
- Models are not always georeferenced. How does that fit with other R2K information?
- Is there any hope to define a common interface, a common format?
- How can we link different levels of abstraction? Who is in charge of that?