search

CSTARS / ecosis

Ecological Spectral Information System (EcoSIS)
4 stars 0 forks source link

Question: Plans for EcoSIS to join DataONE? #46

Open serbinsh opened 4 years ago

serbinsh commented 4 years ago

https://www.dataone.org/

We were recently discussing the archiving of spec+trait data on a data archive, either EcoSIS as primary or another source (Dryad, FigShare) as primary and EcoSIS is the secondary archive; that is the DOI would be that tied to the other source. But then it was suggested that if EcoSIS was part of DataONE then the data would be more discoverable.

Are there plans or thoughts about EcoSIS joining the DataONE alliance?

ptownsend00 commented 4 years ago

Hey Shawn: great question. This had not been on our radar until you raised it, and now it is. Justin and I have gone back and forth over our last two meetings, and are not quite sure, so I thought maybe I'd give you my impressions and look for more feedback. I think for starters we have to identify if we think it is worth it (and is DataOne the only or best option). It would take a significant effort on Justin's part to get ourselves compatible with even the lowest level of DataONE partners (and we would not be able to start this until the first part of next year). As it stands right now, EcoSIS data are pretty shareable either through a google search or via EcoSIS -- which requires knowing about EcoSIS. Is EcoSIS too obscure? We really do want to encourage EcoSIS as the first place to look for spectral data + traits but to maximize people doing that does it require being part of dataone? I really don't want people to chose Dryad instead! I don't know -- any other thoughts?

serbinsh commented 4 years ago

Hey @ptownsend00 thanks for the feedback! OK, perhaps something for a larger discussion but indeed I think the main issue was "discoverability", followed by concerns of maintenance and and end-of-life planning. Let's start with discoverability, because often we have trait data associated with spectra that is useful, obviously, with spectra thus the first thought is to place in EcoSIS, however it could also be useful on its own in other databases or analyses. BUT at present trait data isnt easily found in EcoSIS, particularly among those not thinking about remote sensing issues. As a result we often struggle with whether to place in another archive/db first and then replicate in EcoSIS so that its useful for both trait folks and remote sensing folks. But perhaps this is a slightly different issue than what I raised in the original question? As for obscurity, I think if we keep working at it we can make it less obscure, for sure, and keep citing it in manuscript.

I will say for our DOE-funded work we HAVE to put the data first on the DOE portal(s), get the DOI there, and then replicate the data on EcoSIS. I do this because the data on EcoSIS is far more handy to use because it has a nice API for merging all data together and pulling down a single dataset from the "cloud" in scripts, etc. Most other archives are just that, meaning we have to download and hand-curate on our end - what a pain.

Second, my other concern, though I don't know how warranted, is whether there is a chance EcoSIS would go away and thus all data would be migrated or become hard to find in the future? Its really just about whether if EcoSIS was part of a larger community whether it would be more likely to endure many more funding cycles? This is perhaps a more political question so feel free to ignore answering here.

Happy to setup a call between the main users in our group to discuss thoughts on usage, etc if helpful