Either remove all data from source control, or standardize it

[athewsey]

Currently:

• Some notebooks pull data from the private shared UG bucket (which in some cases has been fixed to point to a public CloudFront URL)
• Some notebooks reference local data files in source control
• Some data files remain in source control that probably aren't referenced anywhere anymore

Regardless of whether the solution is to keep the data in source control, or maintain a public site to download it from, we need consistency and to tidy up any unused assets that could be confusing