Science data life cycle model

From Earth Science Information Partners (ESIP)
Revision as of 12:28, February 13, 2013 by Anne Wilson (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

For our life cycle model, I propose we use the model from the CENDI Report from the Workshop to Improve SDM:

Plan > Collect > Integrate & Transform > Publish > Discover & Inform > Archive or Discard

Here is a pretty picture from the report. (I can't seem to put the actual picture here rather than a link, need more path info?)

One thing lacking in all the models, though maybe it's implied, is a loop. Wouldn't most would agree that a loop back should be made: Discover & Inform > Integrate & Transform? That could cover the versioning and repurposing of data.

Also, I might rename "archive" as "preserve", because "preserve" to me implies an ongoing process, e.g., maintaining access and/or readability.

The model seems comprehensive enough and simple. The names seem to capture most of the aspects I can think of and also take a data rather than a science perspective, e.g., "inform" rather than "analyze". In particular, I like the use of "publish" because I believe there are important issues around that concept.

I've updated my diagram with the life cycle model and taken the liberty to upload it here: Data life cycle activities, concerns, and perspectives.

Please provide feedback!

Follow up:

Via an email discussion we have agreed that this model is sufficient for the time being, with the understanding that "many feedback loops may be interpreted" through the model. From the report that provided the model, these words:

“The science data lifecycle is given in Figure C2. The data outputs of a science project are freestanding artifacts that must be maintained intact, secure, and accessible for future uses, foreseen and unforeseen, but archived or disposed of when no longer useful. Within this linear construct, many data maintenance feedback loops may be interpreted through this simplified model. Programs will plan next set of data acquisitions based on discoveries from the current one and would use lessons learned from data management to plan the evolution of data system for future datasets.”