Numpy: what has changed and what is going to change?

Numpy is used as a data container and its APIs are well-defined. What you might not know about things that have changed lately and what you need to know about changes that are coming

For the past year I have been one of three full-time NumPy contributor paid for by a non-profit organization. We have defined a roadmap, added new interfaces to the package, tightened up some of the APIs, improved performance, and merged a major refactor of the random module.

Why this important, what else is yet to come, and how will this affect you, the end-user.