What Are the Best Data Science MOOCs?
I recently completed a year of self-study through MOOCs with the specific goal of becoming a data scientist. During the course of the past year, I enrolled in and completed 6 online certificate programs on Coursera, Udacity, and edX. Learning from my experience, a great data science MOOC has the following qualities:
- Teaches you practical skills for work as a data scientist like Python and/or R programming, Jupyter Notebooks, scikit-learn, pandas, etc.
- Gives you great feedback for your work.
- Is inexpensive.
- Awards a certificate (Coursera Specialization, Udacity Nanodegree, EdX XSeries ) you can display on LinkedIn and/or your resume.
- Finishes with a capstone project to demonstrate what you’ve learned.
Back in April 2016, I wrote a glowing review of the Coursera Data Science Specializaton by Johns Hopkins University. I still stand by that review, but I was absolutely blown away by the quality of the Machine Learning Engineer Nanodegree on Udacity. This is the best course I have ever taken online or offline. Period. The program is miles ahead of the competition in Data Science education. I was so impressed by the quality and value of the program that I immediately applied for the Artificial Intelligence Engineer Nanodegree the day Udacity announced it. Here are the aspects of the program that really impressed me:
-
50% tuition reimbursement if you complete the program within a year. I completed the Nanodegree in 10 months because I was working on other MOOCs and projects simultaneously (I also took a month long vacation to Europe with my fiancé), ultimiately paying only $1,000 total after reimbursement. It was 100% worth the very reasonable price.
-
All projects are completed using Jupyter Notebooks with Python. This is highly valued by employers. Nearly every presentation at PyData Carolinas 2016 used Jupyter Notebooks.
-
Unbelievably high-quality feedback almost always delivered within 24 hours. You receive full project and code reviews for every assignment in the course and are given a Meet Specifications or Does Not Meet Specifications grade. But don’t worry if your assignment Does Not Meet Specifications the first time – there is no limit on the number of submissions you can make, and every submission will be given a full project and code review with suggestions for improvement. Here is a snapshot of the first page of feedback I received for my final capstone project submission.
-
You get to decide what to work on for your capstone project. I competed in the Kaggle Grupo Bimbo challenge and submitted to my GitHub. I was able to explain my work on this project (and others) during my interviews for data scientist positions. The Nanodegree was an invaluable component of my yearlong data science education and I highly recommend it for anyone aspiring to become a data scientist or machine learning engineer.
