Provide & document access to all datasets used for the examples

[sadielbartholomew]

Thanks to @bnlawrence for noting that there was no documented way to get to the datasets used in the examples.

This PR:

• adds all datasets to a directory: under docs/source which seems most appropriate location-wise);
• adds a zipped version: somewhat tragically, the total size of all the datasets was ~125 MB, which just exceeds the GitHub file size limit, so to avoid the need to use git LFS I have split it into two < 100 MB 'split archive' zipped files;
• sets the tests to use the datasets from the docs/source new location;
• updates the docs 'Examples' page to clarify how to access these datasets in both unzipped and zipped form.