• There isn't really a reason to keep a ton of machinery to propagate
  __version__numpy__ as min numpy supported version throughout the
  repo; we don't do this with other dependencies so I'm sure we can
  manage manually bumping the numpy version at all places when we want
  to.
• Kill Numpy().check() in setupext. Numpy will always be installed due
  to its presence in setup_requires -- unless you don't believe in
  setuptools...
• Simplify the version check at import time. If the module isn't
  present, I think leaving in the ImportError that would have been
  generated is just fine (note that on recent Pythons it's actually a
  ModuleNotFoundError which contains more information, wrapping it in
  an ImportError ourselves actually decreases the amount of information
  available...

PR Summary

PR Checklist

• Has Pytest style unit tests
• Code is Flake 8 compliant
• New features are documented, with examples if plot related
• Documentation is sphinx and numpydoc compliant
• Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
• Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way