Theme 3 Idea, Request for feedback

[hdean83]

I thought of a kind of baseline approach for a Theme 3 notebook building off of the ideas behind the seabird vulnerability notebook. I want to see if anyone has thoughts or tips on how to get it accomplished - and also wanted to check if it just isn't a good idea or maybe it's already being done.

Here's the idea:

A notebook where a person could duplicate it and plug their shapefile in (depicting regions of interest - Essential Fish Habitat, Sea Bird Vulnerability, Marine Protected areas) and then the script would run and provide an assessment of the availability of climate and forecast data from the catalog endpoints.

This would involve a few elements that I'm not sure how to do yet:

• masking a records search using the shapefile
• creating an assessment of the output by number of data sets and variable - maybe in a gridded fashion within each polygon.

I'm thinking of writing the script for Essential Fish Habitat in Alaska. But after it's written, and based on a couple of notebooks I've looked at so far (Working with Vector...; and Geoscience Visualization...) after it's written, a new shapefile could be plugged in at the beginning of the notebook.

[dpsnowden]

I like the idea of breaking up the scenario into smaller pieces that are adaptable to more than one scenario. The seabird scenario is interesting but we need to craft it so that it is more of a system test and less of an AOOS test. By pulling out the extensible pieces we might be able to hit more pieces of the system.

Regarding the shapefile questions. I'd suggest that on the first pass you don't worry too much about an exact record by record match using every polygon in the shapefile and rather focus on the bigger and simpler bounding box for the entire shapefile. Then, if time allows, you could try to create a different script that simply does the polygon search of a catalog in a generic science context.

Start simple and build to more complex scenarios, possibly in separate scripts.

I like the idea though.

[jkupiec]

Making the notebooks more modular also facilitates their use for Baseline tests.

On Mon, May 19, 2014 at 1:14 PM, Derrick Snowden notifications@github.comwrote:

I like the idea of breaking up the scenario into smaller pieces that are adaptable to more than one scenario. The seabird scenario is interesting but we need to craft it so that it is more of a system test and less of an AOOS test. By pulling out the extensible pieces we might be able to hit more pieces of the system.

Regarding the shapefile questions. I'd suggest that on the first pass you don't worry too much about an exact record by record match using every polygon in the shapefile and rather focus on the bigger and simpler bounding box for the entire shapefile. Then, if time allows, you could try to create a different script that simply does the polygon search of a catalog in a generic science context.

Start simple and build to more complex scenarios, possibly in separate scripts.

I like the idea though.

— Reply to this email directly or view it on GitHubhttps://github.com/ioos/system-test/issues/72#issuecomment-43531076 .

[wckoeppen]

I like it. I'll also throw in the counter-argument to Derrick's statement above: the big, simple bounding box would be significantly less useful to how individuals would actually need to use data. But if you're going to take that approach, I would think that most people would want to know which data sets cover all of their polygons, rather than which data sets intersect any part of the larger bounding box.

[hdean83]

Thanks for the feedback! I've started a notebook using essential fish habitat shapefiles as a placeholder. I'm working on figuring out how to turn the polygon layer into a set of bounding boxes on the basis of the geometry of that layer. However, I think there might also be a way to simply mask a search using the polygons. To that end, I listened in on a webinar this afternoon - Catalog Service on the Web (CSW), Semantics and Ontologies (hosted by Atlantic Coastal Zone Information Steering Committee Secretariat) - during which I had the opportunity to ask the group about this issue (maybe others were listening in as well?). I'll make sure to post the video when it's available and links to the presentations in case those might be helpful. The webinar largely focused on dealing with various ontologies and mapping (e.g. MMI).

The group's response to my question about masking CSWs with polygons was that it might actually be complicated to do, but possible. And apparently some CSWs allow masking by polygons (pycsw) while others do not (esri csw). So, creating a script that builds bounding boxes from the coordinates in the layer may be the best bet. The group was not able to provide me with a method during the webinar, so we've followed up via email (Andy Sherin, ACZISC Secretariat and ICAN Steering Group member) - hopefully I'll figure it out soon. Based on Will's input, it seems like this might be a useful tool across a spectrum of interests.