Climate velocity code

[tpoisot]

This is something I've been meaning to add for a little while - this might be a case of porting the code @gottacatchenall wrote for BIAB2?

[gottacatchenall]

The BIAB2 code is a hasty port of some boilerplate from for PCAing a set of layers, with code tacked on to use tiffs specified with input jsons. Should be straightforward enough to implement climate velocity using SimpleSDMDatasets as the source of rasters

[gottacatchenall]

Curious if you have thoughts on chelsa-cmip6. Could be smart for SDT to provide an easy API for custom temporal ranges and specified GCMs. Maybe not for the first climate velocity release, but I think there could be something there

[tpoisot]

Definitely as part of SimpleSDMDatasets -- I haven't read deeply enough to get a good sense of the amount of work required, but I feel like this shouldn't interact with the functions for climate velocity (but it'll make this more useful when it's there)?

[tpoisot]

some boilerplate from for PCAing a set of layers

I think we discussed wrapping MultivariateStats at some point, this might be the time to do it?

[gottacatchenall]

See #204 (comment), we did but it required explicitly calling MultivariateStats.fit to not conflict with StatsBase, which isn't aesthetically wonderful