Data science tools Part 2

Numba – 

This tool is an open source optimizing compiler that uses the LLVM compiler infrastructure to compile Python syntax to machine code.

 The main advantage of working with Numba in data science applications is its speed when 

using code with NumPy arrays since Numba is a NumPy aware compiler. Just like Scikit-

Learn, Numba is also suitable for machine learning applications as its speedups can run 

even faster on hardware that is particularly built for either machine learning or data science applications.

HPAT – 

High-Performance Analytics Toolkit (HPAT) is a compiler-based framework for big data. 

It automatically scales analytics/machine learning codes in Python to bare-metal cluster/cloud 

performance and can optimize specific functions with the @jit decorator.

Cython – 

When working with math-heavy code or code that runs in tight loops, Cython is your best 

choice. 

Cython is a source code translator based on Pyrex that allows you to easily write C 

extensions for Python. What’s more, with the addition of support for integration 

with IPython/Jupyter notebooks, code compiled with Cython can be used in Jupyter 

notebooks via inline annotations just like any other Python code.