Little data, big data, & equipment archival goals & guidelines

I have been thinking about our goals to get all data entered within a few weeks of completion of field season. Also, we have added a new set of animal cams, loggers, etc to our lab inventory.

in-praise-of-little-data

Little data
Goal:
At the end of the 2016 season, I want to take a shot at doing what I proposed in the NSERC grant. Aggregating the little data into big data for the San Joaquin Valley Region from our lab. I see connecting shrub data, subordinate plant data, seedbank data, crickets, leopard lizards, and pollinators as opportunities for connection. I propose we get all little data (individual sampling events and experiments), archived appropriately with meta-data. Each instance is georeferenced and time stamped.  If it is a two-phase structured experiment (shrub-open), perfect. I will then collapse each instance at that level into effect size estimates for comparison. I will not use Rii but instead Hedge’s d or LRR for instance. Alternate approach, secure common measures such as density (abundance of whatever taxa was sampled) and load up micro-environmental estimates (smoothed) and run an SEM.

Worflow:
Use the google drive. Same format as usual.
(1) Cut and paste from primary data files into google sheet to preserve formatting and ensure version control from that point forward.
(2) Tabs for site data, primary data, meta-data, derived estimates such as aggregated means or effect size estimates, and other ancillary data.

141003_-137597960-big-data-icons.-volume-variety-velocity-variability-complexity-f

Big data
Goal:
At the end of this season, by Sept 2015, explore the capacity to mine big data in the form of animal cam pictures (Amanda from Panoche and Taylor from Carrizo Plain National Monument) and pollinator videos (primary Ally).  I see a nice opportunity for some code from Alex, a workflow and semantic development on how to integrate, and the derivation of common measures across all big data in this form. We can also explore bulk processing of all imagery via algorithms or mechanical turks. Output, big data how to paper, primary papers for each study, then a synthesis paper across all big data out there on this topic.

Workflow:
(1) Amanda – archive all animal cam pics on Flickr. Done! yahoo. Use albums, tags, and explore enabling easy mining. Develop clean signal to communicate collaboration and sharing and not my name, ie ecoblender.
(2) Ally – use youtube or the same Flickr account. Put all 2015 videos up asap in albums with tagging emulating decisions by Amanda.
(3) Diego – same approach for pollinator videos.
(4) Taylor – animal cam pics on Flickr for now.
(5) Ally – then begin backloading up older pollinator video libraries asap including alpine sets.

optional_equipment

Equipment
Goal:
Keep track of equipment and preemptive this upcoming season to avoid last minute ordering. Also, align equipment usage and purchasing more effectively with collaborators.

Workflow:
1. Use existing google sheet on drive within ecoblender ‘logistics’ folder entitled ‘ecoblender equipment’.  I copied the 2014 tab and generated a new one for 2015.
2. Collapse previous equipment usage for 2015 into this new doc and include column for grant (NSERC or BLM), principal investigator on equipment, and condition.
3. Sharing and sampling duration are now included as well if we are able to share certain individual items collaboratively.