My boss gave me 30 days (not working days) to learn Python (Data Science). What is the best appr… by Kenneth Goodman
Answer by Kenneth Goodman:
Stanford put out a great Python tutorial with Numpy (Numerical Python/linear algebra library written in C), SciPy (Scientific Python/ Machine Learning library written in C) and MatplotLib (Mathematical Plotting Library), all of these libraries are callable in Python, but are fast because they are running C code under the hood without type checking.
This is a good start, and might take a couple days assuming you know another programming language.
This won’t give you all of the language, but will give you an edge into the field.
This is a more in depth tutorial with good exercises (if the previous one wasn’t enough):
You definitely want to get a grasp on Pandas (library to hold data in tabular format with easy functional interactions):, it also has some stuff on the IPython notebook (IDE in the web, it’s Python’s answer to R, it’s pretty awesome )
Once you are comfortable with everything above, you can move onto a more specific area you/your boss want to focus on.
Some stuff on Neural Nets (deep learning):, , and Google’s course on Udacity:
Some more Data Science style course:. This course is written in R, but you can port of many of the functions, but the course really lays out concepts well.
There is alsoon Udacity, general ML principals.
I know this is a lot, and you DON’T need to do all of it, just pick and choose based on your preference.