Transition into Data Science

I Hate Giving Advice!

  • In this post, I’ve listed things (breadth) you probably need to know to become a data scientist.
  • The depth of each depends totally on the work/job you are learning data science for.
  • Companies are still confused about their job titles in the data-science world. So, do not count on the title. Job description is more important. Two jobs with the same title in the same company and in the same organization can be fundamentally two different things.
  • Keep in mind that almost no job expects you to be a scientist in every one of the topics mentioned in this post.
  • I intentionally did not list basics such as algebra, matrix operation, probability, etc. for two reasons:
    • I expect people with engineering or science background to read this. So, you already have some good knowledge of them.
    • Even if you do not have the background or you do but it is rusty, you can look them up when needed. You will gradually find your way in updating your knowledge when you go deeper into ML/AI material. How about this book?
  • Read/watch/do fast but several times. You should not expect to learn totally new stuff quickly and in one round. Learning takes time but doing can be fast.
  • Limit your learning material. Do not watch many courses or read many text books in parallel. Go step by step.
  • Do not distract yourself with technologies. Python or SQL are not technologies but Tensorflow, Spark, Tableau, etc. are. You can quickly learn them once you have a good grasp of the essentials.
  • Eventually, these are some non-universal suggestions. It may not work for you or it may not fit your learning style.
  • Good Luck!

Start Here

  • Git technology and free hosting websites such as github or gitlab are used for software development and version control. It is a must these days. They ease the process of storing and version control-ing your code. Learn git, create an account on github and make it your routine to keep a copy of whatever you build on github. Those codes will soon become part of your resume. Max 2 days to learn the basics and setup accounts.
  • I used to share this post with people but it is a bit old and extensive and you may get lost. Feel free to take a look, though.
  • This repository is a nice graphical road-map in ML.
  • I suggest reading this report from Workera and check out their website. I particularly like this chart:


Analyst Jobs

Database: SQL Scripting

Data Visualization: Dashboarding

  • At work, you will use dashboards to visualize your findings from data exploration using Tableau, PowerBI, etc –believe me, it is a long list.
  • If you are searching jobs in this category (analyst), consider learning one. I suggest Tableau. Otherwise:
    • some high level familiarity with what they do is enough.
    • you will be able to do what they do in Python.

Generic Data Science Jobs

Machine Learning Foundation

Python for Data Science

Advanced AI Responsibilities

If you are comfortable with the topics mentioned above, you should be able to identify what next advanced topics you want to learn. In this document, I gradually collect references that I found helpful in my own journey. You can find learning material on the topics such as:

  • Deep Learning
  • Vision
  • NLP
  • Reinforcement Learning
  • Advanced Coding
  • Linux bash scripting