Review of OMSA’s MGT6203 – Data Analytics in Business.

My second class in GaTech’s Micromaster program is MGT6203, Data Analytics for Business. Below is my review of the course.

• Quality rating: 1.5/5
• Hours per week: Around 2 - 3 hours
• Difficulty rating: 2/5 (with 5 being very difficult)
• Usefulness rating: 2/5

I had high expectation for the course. I have always been interested in the application of analytics in business settings, and I thought that this is the class where we finally get to learn about it. MGT6203’s course description provided by EdX says that the course “teaches the scientific process of transforming data into insights for making better business decisions. It covers the methodologies, issues, and challenges related to analyzing business data.” It definitely sounded like an awesome course!

Sadly, the course didn’t deliver; it fell way, way below my expectations. What could have been a very interesting application-oriented analytics course turned out to be a very slow, poorly delivered introduction to analytics. The course covers various machine learning techniques through R. Topics we learned included:

• Linear regression
• Logistic regression
• Text analytics
• Graph and social media analytics
• Market basket analysis

It is to be noted that most of these subjects are also covered in more depth in ISYE6501 and other courses. In fact, it sometimes felt as if MGT6203 were simply a poorly-designed subset of ISYE6501.

As I mentioned earlier, the module had a very low workload. The lectures and the accompanying homework were released once every two weeks. It took me around six hours to go through the whole two weeks worth of material, and this was already including doing the optional reading and studying external resources.

I found that some of the lectures, especially earlier lectures on linear and logistic regressions, were quite decent. The lectures introduced each concept slowly, and this slow introduction sometimes allowed for a better understanding of the topics presented. The accompanying multiple-choice homework forced us to apply the concepts to solve problems using R. Thanks to the homework, I definitely gained some proficiency of using R’s tidyverse family of packages – dplyr, tidyr and ggplot2 as well as R’s regression packages.

However, the course began to falter in the second half of the semester. While the course’s slow and superficial-style lent itself rather well when it focuses on easier topics such as regressions, it was ineffective towards learning more advanced topics such as text analysis, network analysis and market basket analysis. These are all very interesting topics, and the course’s touch-and-go treatment didn’t do them justice. In the end, I did not feel like I have gained anything beyond a superficial understanding of these important topics.

The course’s saving grace only came at the very end, when it presented a series of interviews of prominent data scientists, operation researchers and other analytics professionals. The interviews were conducted by the course instructor, Prof Sridhar Narasimhan, and were quite engaging. For non-analytics professionals like myself, these interviews provided an interesting glimpse into the analytics sector and the challenges it entails.

Another thing that I should point out is the final examination, which like the rest of the course was unfortunately poorly designed. The final examination consisted of two parts. The first was a series of 50 multiple choice questions. And oh boy was this bad. Instead of testing students’ understanding about the subject of business analytics, the questions asked the smallest and most trivial details that were mostly glossed over in the lectures. Some of the questions were poorly worded and gave rise to endless confusion.

The second part was an open book coding examination, where we had to write R codes and answer some relatively simple business-related questions. A lot of people struggled with this part as they are not used to code under time pressure and was unable to finish it in time. Fortunately, I did a ton of practice right before the exam and was able to finish and submit my code in the last five minutes. I ended up getting a final score of 92%, which should be enough for me to transfer the MM credits to my OMSA degree (should I get accepted).

All in all, MGT6203 left a lot to be desired. It was a course that could have been, but failed in many aspects. It is unfortunate that MGT6203 is one of the three OMSA classes that make up the Micromaster and was available to the public, as this might reflect badly on the overall quality of OMSA.