Data-Scientist-Learning-Plan

A Data Scientist Learning Plan to not only prepare for interviews, but to become a world class Data Scientist

Now that I've finished Galvanize's Data Science Immersive Program, this is my plan to continue my development as a Data Scientist, not just for interviews, but for my skillset as a whole. But because $$$ is important to live, my short-term focus will be on preparing for the Data Scientist interviews. Once I find the right fit, the plan will shift towards a life-long learning approach towards Data Science.

And in order to continue self-educating myself, a solid plan is very important. I'll be using this Github repository as my central location to store all my practice problems & additional notes.

The plan here will emphasize practice>learning, since I've already finished covering most of the material. For me the focus will be on application of the ideas, so you might notice that the resources are more tailored to this approach. Practicing will help identify holes in my understanding, which optimizes my time spent on filling my weaknesses.

Goal: Find a good-fit between my own skillset and a company's Data Science team.

Requirements:

Get an offer

Pass on-site interviews

Pass take-home examination

Required Skillsets:

Machine Learning Algorithms
SQL
Python
Statistics/Probability

Resources

Codewars(Python): https://Codewars.com

Hacker Rank(Python): https://HackerRank.com

SQLZoo(SQL): https://SQLZoo.net

William C. Chen on Quora(SQL): https://www.quora.com/What-are-some-good-resources-to-practice-SQL-for-a-data-science-interview

William C. Chen on Quora(General Data Science): https://datascienceinterview.quora.com/Answers-1

Galvanize's Immersive Program Assessment Questions(General Data Science)

120 Data Science Interview Questions(General Data Science)

SpringBoard(General Data Science): https://www.springboard.com/blog/data-science-interview-questions/

Andrew Ng(Machine Learning): https://www.coursera.org/learn/machine-learning

Introduction to Statistical Learning(Machine Learning): http://www-bcf.usc.edu/~gareth/ISL/

Introduction to Machine Learning(Machine Learning): http://alex.smola.org/drafts/thebook.pdf

Galvanize Model Comparison

Think Bayes(Statistics): http://www.greenteapress.com/thinkbayes/

Think Stats(Statistics): http://www.greenteapress.com/thinkstats/

Schedule

Since finding a job + developing skills are two separate things, I'm splitting my workday in half. 4 hours for the skills development + 5 hours for networking, applications.

A few principles I'm applying to my schedule:

Morning are the best times I can get my deep work done.
Distributed practice > Mass practice: Psychological research has said that practicing a skill 1 hour over 10 days is better than practicing a skill for 10 hours in 2 days, mainly because of the spacing effect.
Attack areas of weakness first.

Here's my preliminary schedule:

As I continue to interview, I'll identify which skill I'm weakest and that'll get moved to the front of the queue. There's a chance I may need to increase the time on a specific skill due to upcoming interview on a specific topic.

Preliminary Practice Activities

Since I am practicing for interviews, the way I'll be practicing these problems is through solo-whiteboarding, since it most closely resembles how I'll be using the skill. To do this, I'm going to need:

Practice Problems
Solutions to the problems

Here are the initial resources I plan to use:

SQL: SQLZoo.net

Python: Codewars.com , Hackerrank.com

Machine Learning: Feynman Technique on each Machine Learning Algorithms/Interview Questions

Probability/Statistics: Khan Academy + Feynman Technique on Interview Questions

And in addition to all these, working on the take-home practice(s) for one day on the weekends. One day for rest(probably Sunday).

Additional Notes:

I'll attempt the problem without looking, check the answer, apply a memory technique(if necessary), re-do the problem knowing the answer and then move on. Each problem and solution will be logged in this Github repository so I can review previous problems.
For the sake of time, only use a Feynman technique or create a cheat sheet for topics I struggle with. May not have the time to create it for everything.

david-s-he / data-scientist-learning-plan Goto Github PK