there really is a big difference vis. scipy/numpy random sampling

The speed differences between Scipy and Numpy’s probability distributions is immense. Across the board, we see that numpy is almost twice as fast, in all cases. In the worst case, it’s hundreds of times faster. 

I wish scipy would put a warning or something on usage, cause waiting for some of these simulations I actually need to finish will be ridiculous. Thanks, scipy. 

imported from: yetanothergeographer