Just for the record, you can also create full Cocoa-interfaced applications written in Python. XCode used to even include templates for building these applications, but they are gone now from the latest version of XCode[1]
But RaceTripper is right for almost every case: you don't really think about compiling Python[2], but rather it gets interpreted at runtime.
I generally write stand-alone python in SubEthaEdit (substitute your favorite editor here), and only use XCode when I am writing Cocoa-Python applications.
[1] I filed a bug on this with Apple when it happened, but their response was that this was "as expected" and that I could take the templates from older versions of XCode and they would work (this is true).
[2] There are two exceptions to this: 1) There are compilers that will take Python code and translate it into native executables. But people don't often do this in most cases. 2) Python modules are compiled into a bytecode that makes them more efficient to run. But Python takes care of that in the background so you don't generally have to worry about it.
