Decide how to handle large datasets in Scivision

[edwardchalstrey1]

This was a point raised at the REG allocation meeting.

• Could the Scivision frontend support running at scale (in the cloud when run through Binder, say?)
• Might be a good fit for some of the Pangeo tooling
• Something REG could help with

Perhaps to consider during a later phase of the project.

Can use intake for this

[edwardchalstrey1]

Related to #317

[edwardchalstrey1]

One answer: We don't handle "large" datasets that can't fit on Zenodo/GitHub

Perhaps encourage people to upload sample datasets