OMSA Logo

This August 2020, I finally graduated with a Master of Science in Analytics from Georgia Tech. It took me exactly two years over six consecutive semesters, taking six credits (two courses) per semester. It was an amazing journey. There were times I wanted to quit or not put as much effort into an assignment, but I’m glad I could finish saying I did my best and gave it my all.

My Background

My original goal (like probably most people thinking about doing this program) was to help me get into a career in data science. I come from a “non-traditional” background, having worked as a medical device engineer doing mostly mechanical engineering and physical product development. However, I was always interested in the data analysis and statistical storytelling part of my job, and I wanted to get deeper knowledge in those areas. I considered doing a bootcamp but decided to go for a full-on Master’s, because I thought it would give me more credibility, especially since I didn’t have any direct data analytics work experience, and I also thought it would be a great personal achievement. I worked as a medical device engineer for about six years before starting the program.

Why I chose OMSA

When I was choosing where to apply, my criteria was that the school had to have a remote learning option, since I didn’t want to leave my full-time job, and that it had to have a strong reputation. I actually considered and applied to only three programs: UC Berkeley Master of Information and Data Science (MIDS), University of Illinois Master of Computer Science in Data Science (MCS-DS), and Georgia Tech’s Master of Analytics (OMSA). I studied hard and did well enough on my GRE’s, and I made sure to get strong recommendation letters from my current and previous supervisors. However, the decision was made easy for me, because I only got into OMSA. In the end, I’m actually really glad I did OMSA over the other two programs. Here is my reasoning:

  1. Cost: The cost of OMSA is significantly better than MCS-DS and MIDS. OMSA is ~$11K, while MCS-DS is ~$22K, and MIDS is a whopping $70K. Sure, the Georgia Tech name may not carry as far as Berkeley or Illinois in some circles, but in the end, it’s still a Master’s degree from a Top 10 school. Not to mention that GT is the only program that has “Master of Science” on the degree, not some other kind of Master’s. Oh, and rest assured, it won’t have the word “Online” on the degree.
  2. Professors: Sure, the lectures are not live and are pre-recorded for OMSA, but they are taught by the actual professors from the on-campus version of the program, and you can still interact with most of them through the course forums. Great networking opportunity! On the other hand, I’ve heard that some of the teachers for MIDS may not be actual UC Berkeley faculty, but rather working professionals in the field.
  3. History: Georgia Tech already had a great reputation for their Online Master of Computer Science (OMSCS), which was introduced in 2013. Taking that model, GT took their on-campus Master of Analytics program and introduced the online version in 2016. So, it gave me some assurance that they already knew what they were doing with this program, and there would be less technical or administrative hiccups.
  4. Community: The online community for OMSA was fantastic. I made a lot of friends through the Slack community, and people were very friendly and supportive.

You can also read this KDNuggets article that compares the best Online Master of Science in Data Science & Analytics programs.

My Course Journey

There are quite a lot of courses offered in the curriculum. The following describes the courses that I took each semester. I went for the Computational Data Analytics Track:

Fall 2018

ISYE 6501 Introduction to Analytics Modeling

This was my first intro to the OMSA program. I actually took this course as part of the GT/edX MicroMasters in Analytics. I first audited the course back in Summer 2018, and I was worried, because it seemed way too difficult. Thankfully, when you actually take it as part of the MicroMasters or OMSA, you can get a lot of help from TA’s and fellow students through Piazza forums and Slack. Plus, you can attend office hours, which are very helpful in completing the assignments. I found the quizzes and tests quite difficult and required a good amount of study. Also, the course used a lot of R, which I had to learn a lot on the fly. I definitely recommend brushing up on R before taking this course. The lectures are given by Dr. Joel Sokol, and they are very good. I liked that he injected a bit of humor into his lectures. I often would refer back to the lectures and notes from this course to prepare for job interviews.

CSE 6040 Introduction to Computing for Data Analysis

This was one of my favorite courses. It was entirely focused on using Python for data analysis. Yes, I spent many hours and nights debugging and getting my code to do what it wanted, but it was at the same time challenging and just doable for me to complete the assignments and exams. The course is run very well, and it improved my Python significantly. It was also my first introduction to pandas, which I would use a lot later on in the course and in my career.

Spring 2019

CS 6400 Database Systems: Concept & Design

This was a very interesting course. It was the first course I took that was shared with students from the OMSCS program, since it is offered as part of their curriculum as well. It involved a group project, and I was lucky enough to be in an awesome group. I found that, not surprisingly, the OMSCS students usually had more coding experience than OMSA students. The group project involved building a database warehouse starting from database schema design with EER diagrams to an interactive user interface. This course helped improve my SQL significantly as well as learn some best practices for designing database systems.

MGT 6203 Introduction to Business for Data Analytics

People tend to hate on this course, but I honestly don’t understand why. I heard it has been redone since I took it. I found the homeworks very helpful to apply the concepts in R. The course had less to do with business than it had to do with data analytics. The most interesting portion of the course was when it went into NLP with text sentiment analysis and LDA. It was definitely one of the easier and less time-consuming courses, though.

Summer 2019

ISYE 6414 Regression Analysis

This is another course that doesn’t seem to have the best reputation, but I thought it was great. Linear Regression is still one of the most widely used “machine learning” techniques, yet there are still a lot of incorrect assumptions about regression. The course also gets into ANOVA, logistic regression, nonparametric regression, and error metrics. The quizzes and exams were quite difficult and contained some “gotcha” questions. The course does go into regularization (ridge regression, LASSO, elastic net), which is very useful and practical in real life.

MGT 8803/6754 Business Fundamentals for Analytics

This was the most different and unique course. It is a required course, and it has been referred to as a mini MBA in a class. I actually think it was smart of GT to require this course in the curriculum, because I’m finding that Data Scientists need to be business-savvy and understand how companies are financed. Some topics included Financial Accounting, Managerial Accounting, Business Strategy, and Entrepreneurial Finance. Yes, it was boring at times for someone like me who’d rather do more data analysis. However, I think it helped me understand the right business questions to ask, and how to evaluate and value companies.

Fall 2019

CSE 6242 Data & Visual Analytics

This was one of the most difficult courses for me. It was very fast-paced, and each of the 4 large HW assignments required at least 10 hours of work. On top of that, there is a group project, where you have to develop something “innovative” using a “large” dataset. There was also a lot of data engineering and ETL taught in this course, and it exposed me to a wide variety of tools, such as D3.js, AWS, Azure, Hadoop, Pig, Spark, Scala, and API’s. It definitely helped improve my coding and ability to learn things quickly.

ISYE 6644 Simulation

I found the lectures for this course to be among the best. Professor Goldsman is a joy to watch and is very entertaining, which is tough to do for a statistics course. I was a little rusty on calculus and probability distributions, so this course helped to refresh my memory and go deeper into PDF’s, CDF’s, the means and variances for different distributions, Monte Carlo simulations, and much more. A good portion of this course focused on using the software Arena, which I’m not sure how widely it is used in practice, but I could see it being really useful for manufacturing facilities, call centers, etc.

Spring 2020

ISYE 6420 Bayesian Statistics

I had always heard that it is important for Data Scientists to be familiar with the Bayesian formula, but I didn’t really know what that meant practically. That’s why I decided to take this course. The first part of this course went pretty heavy into probability and distributions, and I’m glad that I took ISYE 6644 Simulation beforehand. In the later part of the course, we used a software called OpenBUGS to go through Bayesian examples and had to write our own code to solve Bayesian-type problems.

ISYE 6740 Computational Data Analysis / Machine Learning

This was probably my favorite course in the program. I could finally put together what I had learned from previous courses to go deeper into the math behind popular machine learning algorithms, and to use popular libraries like scikit-learn to program and solve machine learning problems. Anybody can call a package and apply a machine learning algorithm to a dataset, but I think it is important for any Data Scientist or Machine Learning Engineer to really understand what is going on under the hood, in order to improve or extend predictions, to explain how the algorithms work, knowing when to use a specific algorithm, and to understand the pros and cons between different algorithms. The course was so interesting to me that I joined the course as a TA during the Summer 2020 semester and continue to work as an Instructional Associate after I graduated.

Summer 2020

CSE 6748 Applied Analytics Practicum

Some people choose to do this project earlier, but I’m glad that I chose to end with this course, since it involves working with a real company to solve analytics problems. You have the option of working on a project with your current employer, or choose one of the projects that GT provides with one of their sponsored companies. Since I was trying to make a career transition, I opted for the GT-sponsored project. For my project, I worked with the marketing department of a software testing company and used data analytics to derive insights and improve their marketing lead to sales conversion. The biggest thing I learned from this project was dealing with “dirty” data. I had always heard that data wrangling is usually the most time-consuming step of a data science project, and I got to experience that first-hand with this project. Most homework assignments given in my previous coursework used data that was already pretty clean, so I rarely had to clean the data myself. This course was also a great resume-booster, since it helped me gain some related work experience that would otherwise have been difficult to get. The practicum is also a distinguishing factor between OMSA and OMSCS, since OMSCS currently does not offer a practicum.

There are other courses offered in the curriculum, and GT is constantly adding more courses. Some courses that I didn’t get to take that looked interesting are CSE 6250 Big Data Analytics in Healthcare, ISYE 8803 High Dimensional Data Analytics, and Deep Learning (will be offered soon). For supplemental learning, GT offered many resources, such as free access to all of the O’Reilly books, LinkedIn Learning, Codecademy Pro, MATLAB, PyCharm Professional, Office 365, and many others. I would definitely recommend taking advantage of all of the student perks.

Conclusion

For anyone considering Georgia Tech’s Online Master of Science in Analytics (OMSA) program, I would highly recommend it. I went from not knowing what pandas was to being able to wrangle complex data sets without referencing Stack Overflow too much. Before the program, I didn’t really know what machine learning was, and now I can describe the underlying math behind most popular ML algorithms, describe the pros and cons of each, and know when to apply them.

There were all types of people at different stages in their career. Some, like me, were coming from a non-traditional background and trying to transition into data science. Some were already data analysts or scientists and just wanted to deepen their understanding. Some were seasoned, older professionals that were just doing the program for fun. I even heard that the youngest person enrolled in the program was 16 years old!

Bottom line: The GT OMSA program was definitely worth it. For me, it was a no-brainer. I was trying to make a career transition, already had basic probability and statistics and programming knowledge, and I didn’t want to sacrifice my full-time job. My previous employer reimbursed my tuition (although it was not much anyway). Although there’s still always much to learn, and the field is constantly evolving, I feel confident and well-equipped to begin my career in Data Science, thanks to Georgia Tech’s Master of Analytics program!