Review of OMSA’s ISYE6501 - Introduction to Analytics Modeling

To prepare for my application for the OMSA in June, I am now completing Georgia Tech’s Micromaster in Analytics. This mini-program, hosted by EdX, consists of three basic courses in programming, business analytics and statistics and serves as a teaser as to how the OMSA is like. If you score well in these courses (>85%) and are admitted into the OMSA, credits from the Micromaster will be counted towards the graduation requirement. Considering that the per-credit course of Micromaster is cheaper than the per-credit course of OMSA, completing Micromaster can also help you reduce your overall tuition fees.

The first course I am taking is ISYE6501, Introduction to Analytics Modeling. Below is my review of the course.

• Quality rating:4.5/5
• Hours per week: Around 10 hours
• Difficulty rating: 3/5 (with 5 being very difficult)
• Usefulness rating: 5/5

ISYE6501x is designed as a ‘teaser’ course – one that gives you a broad picture of the field of data analytics. It takes you to a whirlwind tour of the most common techniques used in analytics, including the following:

• 'Standard' machine learning algorithm, such as linear and logistics regression, kNN-clustering and SVM
• Basics of optimization
• Basics of simulation
• A quick review of some advanced concepts of statistics and probability

All in all, I thought the course was very well designed. In fact, taking this course was one of the factors that convinced me of the quality of GaTech’s OMSA program. The course really achieves what is meant for – to introduce students to the world of analytics and expose them to some commonly used analytical tools. Dr Joel Sokol, the course’s instructor who happens to be OMSA’s director, is a great teacher who can communicate challenging concepts in succinct and sometimes humorous way.

The course covers a lot of topics; however, due to time constraints, the lectures do not go deep into the mathematical underpinnings of the models. For instance, when we discussed about PCA, the course did not provide a step-by-step explanation on how to obtain PCAs from a data matrix, which is a more common approach in a linear algebra course. Having said that, the course does a good job in ensuring that students understand (1) when to deploy a certain analytical model, (2) the advantages and disadvantages of each model and (3) how to construct the model in R. As it is a broad survey course, I believe that anyone considering to enroll in OMSA should read ISYE6501 as their first course.

I did not find the course difficult at all. The first two weeks were a bit challenging for me because I was still getting used to coding in R. However, as I became familiar with R, it definitely got better. I also needed to get up to speed quickly with some basic probability/statistics concepts that I had forgotten (such as the various statistical tests and probability distributions), but this was not too difficult. I recommend anyone who want to take the course to first familiarize themselves with R, probability and statistics.

The main component of the course’s assignment is the weekly homeworks, the bulk of which consists of R coding. The questions are rather rudimentary and the programming office hours provide you with clear pointers to tackle them. These homeworks are peer-reviewed. Due to the TAs’s constant reminder for us to be lenient in grading our peers, if you really put in enough effort, you will almost sure to get 100% for your homework.

Another component is the mid-terms and exam, which consist of multiple choice questions. The tests are very straightforward and do an excellent job of testing students’ high level understanding of the concepts. No live coding or complex computation was required, and most students finish the exams within half of the time allotted.

I took the course in the condensed summer semester and I still thought that the coursework was quite light. I spent around 10 hours each week – watching lectures, reading the ISLR book (an excellent accompaniment to the course and a must-read for any aspiring data scientists) and doing the homeworks. I completed the course with 92% – a decent score that I am sure have helped me in my OMSA application.