#DataScienceResources

[ashutosh4]

Learn about the architectures, best practices, and design concepts of AWS services. cloud computing know-how for robust and scalable solutions. #AWSExpert #CloudProfessional

For more information visit www.certera.co

[makoyandcheese-blog]

Study Plan (Part 1)

First thing I did when I decided to be more familiar with data science is to talk to someone in the field. 

When I started working, James* and I are in the same boot camp which equips us with the basic skills and knowledge before we are given projects. For some reasons he left the company not long after the camp and we kept in touch from time to time. We usually talk about Python programming and anything related to data science and analysis. That’s when I asked him what are the first step in studying data science. He told me to have a strong mathematical and statistical foundation and comfortable in programming (Python or R). Some of the topics that I must focus is Linear Algebra and Multi-variable calculus. He also suggested that I join Data Science groups in Facebook and go to meetups.

And so I listed topics that I’ll be studying and resources that I’ll be using. I have found samples of learning plans online and these gave me a rough idea on how I approach my creation of a personal study plan.

For Math, I have taken a number of units in college due to my major. Some of the subjects I have taken is precalculus, Calculus for Computer Science and Discrete Math. If specializing in Data Science, we will be required to take Statistics but since I didn’t I might need to focus on that. And I badly need a refresher since I’m not sure if we’ve really touched the topics of Linear Algebra (LA) and Multi-variable calculus (MVC). I used Khan Academy videos (Linear Algebra & Multi-variable Calculus) as a starter. The MIT Open Courseware also has a topic for LA (here) and MVC (here) which I will check once I’m done with Khan Academy though some sections may overlap, a real lecture from university won’t hurt. It seems that they have a number of lectures for a single topic so maybe I’ll use the recent one. Analytics Vidhya also has a helpful guide (here) for beginners and I personally like how it started by asking “Why study Linear Algebra?” and why it is important in learning Data Science.

And then there’s Statistics. From what I have gathered, topics that I need to focus on is Probability, Descriptive and Inferential Statistics. Udacity has some great free courses that will give me an intro to Stat. I’m currently taking their Intro to Statistics (here) and it was a pretty smooth ride until I got to Bayes’ Theorem which had me rewinding the videos to understand it properly. They also have an Intro to Descriptive (here) and Inferential (here) statistics. Phew, I have lots of stuff to study and to squeeze in between work schedule. For probability, edX has a course “ Probability - The Science of Uncertainty and Data“ but one if its pre-requisites is Multi-variable. There are other courses related to Statistics that I will check in the future. As of now, I’ll stick with the above.

Now for programming, I have a background in Python (I mainly use Java in college) and will just need to regularly practice to get used to it again. Well, there’s the staple Codecademy but I might just do challenges in Code Signal. I regularly use the latter (formerly known as CodeFights) in college to continuously refine my programming and problem solving skills. I think I learn best when challenged with a problem. And there’s SQL and Udacity has a course called SQL for Data Analysis (here) but I haven’t check it out thoroughly.

As of now, I’m currently brushing up with my Math Foundation while looking for more resources for other topics such Python Libraries, Data Visualization, Data Cleaning, Data Mining (do I need this?), and Data Analysis. I have seen some resources for Machine Learning, Deep Learning, and Neural Network though I need to check these further and look more. Also, if I’ll be looking for a job, almost all of them are looking for a portfolio and I have no idea how to create one. I will need to dig deeper on this and look for public data sets which I can experiment on.