Test using watersheds to group parks

[danrademacher]

To help improve issues around radial distance vs true travel time, where 9 miles as the crow flies can mean an hour's drive, Seth's going to spend no more than 4 hours testing a simple method of using watersheds to push "in-basin" parks to the top of search results.

[mojodna]

This is what the HUC-8s look like:

It would address the Santa Rosa / Napa adjacency, but would behave confusingly with SF as a source unless we did something like find all watersheds within threshold.

[danrademacher]

Is the threshold option easily done? What do LA and San Diego look like?

[mojodna]

LA:

San Diego:

[mojodna]

Sacramento:

[danrademacher]

So SF, la, and sacto would all benefit from the threshold. Hmm

[danrademacher]

What do you think we should do?

[mojodna]

I'm going to sleep on it and give it a whirl in the morning. I'm about 20 min deep so far. I think it'll work.

[mojodna]

An interesting approach, but ultimately flawed for the Bothe-Napa Valley case because that sits in both watersheds (and thus probably others that matter):

I'm inclined to keep things simple by not introducing this and hoping that we can do something with driving distance in the future.