Learning Datascience/Machine Learning

Do you want to get started as a datascientist? I can help you!

• I have a strong background in machine learning theory
• I am familiar with Python's datascience toolset (numpy, pandas, sklearn/scikit-learn, matplotlib, seaborn, jupyter, ipython, pycharm, keras)
• Have experience with kaggle competitions, kaggle kernels, and how to use them to learn python for data science.
• Have experience with the fast.ai deep learning course
• Am connected with the data science community in Vancouver

How to get started - Get the big picture and get connected:

1) Start with Andrew Ng's machine learning coursera course (A $200US/month alternative is Udacity's Machine Learning Nano-degree)

2) Come to the Learn Data Science by Doing Kaggle Competitions meetup, and to the pub afterward

3) Consider the fast.ai course if you are interested in deep learning (very hot right now). This free course gets you hands-on right away (top down).

4) Consider the deeplearning.ai course for a bottom-up introduction to deep learning.

Once you've gotten started or run into difficulties, I can help you!

Other Study options

To brush up on your math - You just need to understand derivatives in the single and multi-variable cases, and basic matrix multiplication is occasionally useful:

• Khan academy on differential calculus AND/OR
• A MOOC on differential calculus
• Start a MOOC on Linear Algebra (Matrix multiplication is occasionally useful)

To brush up on your python:

• The python tutorial is good
• Then I would learn what you need as you go (i.e. fo

Practice on kaggle competitions

• The titanic competition is a good option to start with
• Reproducing kaggle kernels (like this or this) from scratch is a good way to pick up the Python syntax and workflow you will need.

Other structured programs (Which I'm less familiar with, but may be as good/better than udacity/coursera)

• https://www.datacamp.com/ - choose Python, not R
• https://www.springboard.com/workshops/data-science - Unfortunately this starts with R, not Python