Hey Ryan. So the absolute best "non-standard" best practice I can suggest is to NOT establish a list of "Best Practices." It tends to stifle creativity, innovation and after a time can feel very oppressive. It turns into a paint-by-numbers formula that can be completely divorced from the sea/landscape it purports to represent, or the data you're trying to visualize.
Instead make two list:
- Things that normally seem to work
- Things that normally tend not to work
"Seem" and "Tend" are important, because then we get away from "Always" and "Never", which are the downfall of all "Best Practices" lists.