Reflecting on the Data Science Certificate in Coursera

Throughout the beginning of the year, I kept getting ads in my Facebook feed about a Data Science Certificate offered by Johns Hopkins University Blooomberg School of Public Health on the Coursera platform.  I hardly ever click on ads in social media feeds, but give credit where it's due: someone had their target marketing campaign dialed in correctly, because I clicked on it.  Score one for their conversion model.

I reviewed the curriculum and found it interesting; data science is such a hot topic right now, and I wanted to learn more about it with some hands on experience to boot.  This seemed like a good way to get my hands dirty, but the price of the certificate and the payment model (pay-as-you-go on a per-course basis) lowered the barrier to entry for signing up.  

I had never taken a course online until this year, but the Coursera platform is excellent for online learning.  The positives: Students were actively engaged in the online forums and the Community TAs were very attentive.  The content and lectures were well assembled for self paced study.  A word of caution: I could do this certificate in my spare time because I had a lot of experience working in databases and software development. Having good fundamental computer science skills made the programming projects easier.  If you are considering this certificate, and you struggle with code, either allow more time for the assigned projects or perhaps look for some free programming courses available in Coursera.

The best part of certificate was learning to program in R and refreshing my math and statistics skills.  Some of the subject matter could have been overwhelming if not for practical exercises in R.   I was pleasantly surprised how much better I understood statistical concepts when they were taught at an intelligent level in the context of a programming language.  Perhaps my brain is just wired that way, but if I had simply been taking notes and writing formulas with a bunch of Greek letters, the concepts would not have stuck as well.  While Stat Inference and Regression were the most difficult courses for me personally, I am thankful that the math was taught at levels deeper than I got in my MBA statistics courses.  (I'm also a musician, and I have had the same experience in learning music theory.  Things become more fun when one has a deeper understanding.)

I've already realized the benefits of the program.  For example, I can speak with greater understanding about stats and regression with my boss (who has a Ph.D.) , using R has made regression analyses faster than doing the same analysis in Excel (plus using R and other code sharing technologies such as Git makes my work reproducible).  Finally, I can better frame questions about data and guide an analysis rather than making anecdotal observations.

I'm very grateful for all that I learned in 2015 (In the spring, I will complete the capstone project which draws from all of the courses in program).   I hope that 2016 will afford opportunities to flex these new skills. It's an amazing time to be around data science  not only for the tools we have to do the work (R, Git, R Studio, Shiny, etc.), but also for the powerfully flexible tools we have for learning.