How to build a cldr.json file as in the globalize compiler example

I am finally looking to migrate from 0.x to a current version of globalize and I admit I have a very hard time doing that. I apologize if I ask stupid questions.

So before, I only had to load the desired culture file and work with it - but now, I have loads of CLDR data and I do not need the most of that.

So first of all, I want to know how to build a custom cldr.json file like provided in the compiler example, so I wont have to load >500KB from lots of separate files like in the other examples, but only the actual data I need. Right now I only need basic Number and Date parsing and formatting .

So, how was this file actually built? Maybe I have overlooked something, but I could not find any hints in the docs. Plus, is there a way doing this using grunt? The ideal way of doing this would be using an automated task where I can specify the desired locale(s) and excactly which kind of data I want to be included, resulting in a compact file like in the example.

Any help is greatly appreciated!