Dask & Numba: Simple libraries for optimizing scientific python code

Python is a high level language that is used by scientists for numeric computations. However, the performance of the language can be a hindrance when scaling to larger data sets, requiring some operations to be rewritten in a lower level language. To address this problem, we propose two libraries to allow numeric Python code to be optimized incrementally, requiring minimal changes. Here we describe Numba, a compiler for a subset of the Python language, and Dask, a flexible parallel programming library.