MOOCs, Data Science, and me

Last update: Friday 26 January 2018

This post is a personal note from the DLL Editor, a belated follow-up to previous posts that have marked his progress via MOOCs towards becoming a data scientist ... :-)

Readers can offer me their congratulations and/or condolences on my completion of another online Data Science certificate program -- this one from DataCamp (23 short courses); my first online DS certificate was from Johns Hopkins U via Coursera (9 courses plus a capstone course). 


I am now within a capstone of completing a certificate program for "Statistics with R" from Duke U via Coursera (five courses) and another certificate program for "Advanced R Programming" from Johns Hopkins U via Coursera (five courses). But I still don't regard myself as a "Data Scientist" ... yet.

My working definition of what a new data scientist should know is what I think I might have learned had I been able to enroll in a rigorous, two year, full time Masters degree program. 

The lingua franca of the leading publications of the statistical branch of data science which I am studying (compared to the computer science branch) is calculus plus linear algebra. Not having used either subject during the last 50 years, I found a course in linear algebra (offered by UT Austin via edX) that starts next week. But the good news is that the calculus courses from U of Penn via Coursera that I started last week have quickly proven to be the most enjoyable of all of the many online courses I have taken in the last three years; indeed, they should prove to be some of the most enjoyable (and hardest) courses I have ever taken.

I was pleasantly surprised by how much calculus I remembered last semester while coaching my granddaughter through an intro calculus course she was taking at U of Maryland. Then I looked for rigorous review courses for folks who had taken calculus a "long time ago" ... didn't quite find that, but found a close equivalent. The courses I am now taking from Prof Ghrist at U of Penn are described as intro courses for undergrads who did well in their AP calculus courses in high school. In other words the courses assume that students already know what limits, derivatives, integrals, infinite series, etc., are all about. Professor Ghrist's goal is to teach students what these concepts really mean. Here's a quote from the introduction to the first course:

• "Calculus is one of the grandest achievements of human thought, explaining everything from planetary orbits to the optimal size of a city to the periodicity of a heartbeat ... Distinguishing features of the course include: 1) the introduction and use of Taylor series and approximations from the beginning; 2) a novel synthesis of discrete and continuous forms of Calculus; 3) an emphasis on the conceptual over the computational; and 4) a clear, dynamic, unified approach."

As a recently retired academic, I have been pained by the intensive resistance from a powerful rear guard of the professoriate who are determined to block further expansion of online courses in the for-credit curriculum.  IMHO this is a losing war fought by folk who misread the significance of MOOCs.  I suspect that much of their resistance to online courses stems from their never having taken a really great online course themselves, i.e., a course that caused profound tectonic shifts in their intellectual perspectives. So they implicitly and/or explicitly compare the many mediocre online courses that they have "examined" to the very small handful of excellent face-to-face courses that they themselves took as undergrads or grad students.

The best face-to-face courses I ever took were taught by professors who were convinced that all previous courses in the subject, however well-intended, missed the mark. Soooooo they set out to do it "right". As you can tell from the quote from Professor Ghrist, he also set out to design yet another intro to calculus because all of the previous courses hadn't quite done the job. So far, I think he has hit the high marks he set for himself. It's been 60 years since I was first stunned by Euler's formula and by the notion that any "reasonable" function can be represented by a Taylor series composed of data obtained from the derivatives of that function at a single point, x = 0.  Prof. Ghrist hit both of those power gongs in the first two weeks!!!

Unfortunately I can't recommend these courses to the dwindling academic rear guard because Professor Ghrist deliberately set out to teach "hard" courses that have extensive prerequisites.  I have read about other great online courses in other fields that might be more accessible to the critics. But as in the face-to-face world, great online courses are still few and far between. So I suspect that the "good guys" won't win the "Online Wars" for another 15 to 20 years, the time required for a generation of academics who took two or three great online courses as students to attain tenure and the power to outvote the rear guard on curriculum committees ... :-)

Roy L Beasley, PhD

DLL Editor

++++++++++++++++

Related notes on this blog:

• Another "Masters" in Data Science via MOOCs ... January 2018
   
• Directory of Academic Degrees & Certificate Programs in Data Science ... preliminary draft ... January 2018 
   
• Job-Oriented Online Programs (JOOPs) ... August 2017