“… decisions to archive data are currently made by individual researchers, and it is [not] obvious that the benefits of Public Data Archiving outweigh the costs for all individuals.”
- Roche et al. 2014
What does it mean to share data publicly?
Your data will be available for anyone, to do anything
Creative Commons Zero means open! but… you have no claim to prevent anyone from altering your data, no right to help with interpretation, or to deny someone from misusing your data.
Your data must be in an open format
- there are recommended formats, e.g. Dryad recommended filetypes
- should be a standard format
- should be a non-commercial format
- should be a human-readable format where possible
Your data will be available consistently in perpetuity
- Digital Object Identifier e.g., try searching “10.5061/dryad.6t94p”
You should include
- data
- metadata
- code
- README and other documentation necessary to understand structure
What are the main reasons to make your data publicly useable?
Group fitness
The broader research community, and the general public, can benefit from the most data available to answer a particular question of interest, especially those for which the data were not intended.
Reproducibility, a key topic (and a foundation of the trust given to scientific results), is facilitated by data availability.
Individual fitness
- Grant money: many agencies requiring release, at least nominally
- Altmetrics
- Data Usage Index
- Data citation
- Fame, if not fortune…
But what about the impact of the shared data on your future productivity? Will sharing data now make you “less productive,” even if it makes the overall research community more productive?
What are the main places to make your data publicly useable?
In general, the objectives are to have a stable, long-term place for the data, as well as have the data be indexed and citeable.
In conjunction with publication
Free-standing or in conjunction with work in progress
- Figshare
- upload, get DOI, share (private or public)
- embargo up to 2 years
- CC By 4.0 or CC-0
- Zenodo
- upload, get DOI, share (private or public)
- CC By 4.0
- Github
- e.g., PortalData
- Environmental Data Initiative
- CC0
- allows embargo and “offline” private sharing
- archived & indexed by DataOne
- LTER data standard
- DIY data service
- e.g., Datasette
- web or blog hosting
- immediately useful, less traceable or citeable
- others: search re3data.org for domain-specific registries
Archiving: for the long term
-
Dryad + DataOne partnership