And remember guys, less is more :)
I agree with the premise here, but the first example displayed contains far more complex data than the follow-up examples. It's seems trivial to make a nice dataset easy to read, but how about the first example? That's a tough dataset to work with, and I feel like the rules suggested don't really go very far to helping to make that table more useful and legible, beyond just splitting into smaller tables that have simple labelling/row/column systems.
Less IS more. I wish this was used more in financial reports.