Persist datasets locally?

[sckott]

I wonder if it makes sense to persist datasets on a user's machine. And then read the cached version unless user wants to update, or possibly check if any changes on remote dataset

using rappdirs maybe

[jeroen]

That might make sense, although the user could easily do that herself by first cloning the data package directory and then opening the pkg from disk.

[sckott]

And won't make much diff. for small datasets

[richfitz]

This is literally what I do (incl with rappdirs) with datastorr. I even have a datapkg issue there (https://github.com/richfitz/datastorr/issues/2