An honest, post-graduation review of the course’s layout, syllabus, and instructors.
General Assembly bills itself as a leader in Education. Founded in 2011, GA claims to have “transformed tens of thousands of careers through pioneering, experiential education in today’s most in-demand skills”1 such as data science, web development, and user-experience design. GA offers a wide range of classes and workshops, from 2-hour info sessions to 3-month, full time immersion courses.
My first introduction to General Assembly was in early 2017 when I enrolled in their Excel and SQL bootcamps. These workshops were either one (Excel) or two-day (SQL) accelerated overviews of the respective tools.
After six hours in the Excel bootcamp, I actually was able to walk away with a much better understanding of conditional statements, formulas, and pivot tables and was able to implement the techniques I learned in the class to my job immediately. The SQL class also helped, but I wasn’t able to practice writing queries as much after the class.
But as the months went by, I had a lingering feeling that if I wanted to advance my career in any kind of meaningful way, I needed harder-to-obtain skills than knowing how to write some formulas or manipulate some pivot tables. In the year prior up to enrolling in GA’s Data Science Immersive course, I kept returning to the curriculum in which GA promised to teach its students the fundamentals of machine learning in 12 weeks. I was frustrated with how stagnant and limited my skillset was at my current job and concerned with my ability to advance beyond a lower-mid level role. I wanted to take on more interesting (and analytical) projects but was routinely overlooked in favor of someone with more technical skills.
For about a year, I kept returning to that curriculum, nervous about how I could possibly swing taking three (plus) months off of work while living in NYC, and more nervous about what would happen if I didn’t.
Orientation Day
What GA (and any other bootcamp for that matter) proposes to accomplish is somewhat of an impossible task: Take students from varying backgrounds, skillsets, and aptitudes and teach a wide breadth of new and complicated material in what ends up being a sprint to the finish line every single week. My cohort was no different. At one end of the academic spectrum, a student decided to matriculate through the bootcamp in lieu of going to college. On the other, there was a student who had tripled majored in math, physics, and philosophy at Yale. There were students who had had professional experience as data analysts and were looking to advance their careers in their established fields and there were those that were looking to transition from completely unrelated industries. The majority of the class had their Bachelor’s degree, but a healthy amount also had Master’s. One student had a Ph.D in Statistics. About half of the class had majored in Math or a hard Science. This is all to say, while it was a diverse cohort, many of my fellow students were quantitatively inclined.
Syllabus
The Data Science course is taught simultaneously across the U.S. divided into two sections - East Coast and West Coast using a “connected classroom” format. Using a video call, a global instructor would teach the lecture material to all five of the East locations (New York, Boston, DC, Atlanta, and Austin) and then turn around and teach the same material to the West Coast cohort.
Surprisingly, it was fairly easy to engage with the lecturer via a dedicated Slack channel. During lecture, students would pose questions on Slack, and either the global instructor would answer them on the spot or, more often, lecture would continue while the local instructors, who were in the classroom with us, would answer them via Slack. If you think this sounds like there was a lot happening all of the time during lecture, you’re right. There were multiple topic-related side discussions going on at once during lecture that you could choose to follow. (They were almost always related to a spin-off question posed during lecture.) This was my first time using Slack, but I soon learned to love it because it allowed me to go back and review a written record of the lecture, and specifically, answers to questions that I may not have fully understood the first time around.
I included a diagram of the schedule (seen below) from GA’s website because it more or less describes how class was run every day. (NB: While the course material is the same across the different locations, the overall experience of the course varies slightly based on the local instructor, the makeup of the class, and class-time restrictions mandated at the state level. Because I am based in New York and took the course in New York, this post only describes my experience in that location.)
I would caveat that while the post-lecture reviews are technically optional, they really aren’t in practice. This is time that the local instructor uses to go over questions related to the lecture material or work with students individually to answer questions related to labs or projects. Trust me, you won’t want to skip out on this. In New York, there wasn’t a “Community Meetup” that I can recall. We basically just ended lecture around 4 PM, took a ten minute break, and then had our afternoon lab/review time right up until the end of the day at 6 PM when our local instructor congratulated us for surviving another day.
While I’m sure the material changes slightly with each cohort, this is a summary of the topics we covered each week:
Week 1: Intro to Python
Week 2: “Pandas Appreciation Week”, SQL
Week 3: Linear Regression
Week 4: Classification
Week 5: Web-scraping, APIs, Natural Language Processing
Week 6: Advanced Modeling
Week 7: Neural Networks
Week 8: Unsupervised Learning
Week 9: Bayesian Statistics
Week 10: Spatio-temporal Modeling
Week 11: Big Data (AWS, Scala, Spark)
Week 12: Capstone Project Workshops
Overall, I was happy with the breadth of topics we covered. As with any bootcamp, depth is always going to be a challenge given how much material there is to cover within a certain amount of time. That being said, GA provided many resources to further explore any given topic, and post-course, I have found myself reviewing materials and lesson plans.
There is, however, one change I would make to the syllabus: Week 9. We spent a whole week on high-concept Bayesian statistics, and given that I majored in Ancient Greek, this was an uphill battle. While I do believe that a fundamental grounding in statistical analysis is important to Data Science, I believe our time would have been better spent learning R - another highly sought-after programming skill geared more toward statistical analysis. When time permits, that will be my next challenge.
Meet the Instructors
There were four global instructors, all with impressive credentials in statistics, machine learning, and development. Three out of the four lecturers were excellent. They were able to explain difficult concepts in an organized and coherent way, often with supplemental powerpoint materials that enhanced my ability to grasp new concepts, especially ones that were more math and theory-based. There were also many in-class “code-alongs” during which we practiced actual execution and wrote code in tandem with the instructor to reinforce programming skills. This technique was particularly useful when we were first learning how to use the neural network libraries, Keras and TensorFlow.
We also had a local instructor and a TA in class with us. I cannot say enough good things about my local instructor. He had the impossible task of raising 20 baby Data Scientists and did it with the patience of a demi-god. He explained and reviewed difficult concepts multiple times until things began to stick. He debugged coding errors like a ninja. He supported my ideas for projects and my capstone and he helped me achieve my overall goals for the course. He was approachable, likable, and compassionate and one of the main reasons I was able to make it through 12 weeks of Data Science.
Graduation Day
We also had a weekly Outcomes meeting every Thursday afternoon during which we discussed topics related to our pending job searches.
In addition to our weekly homework assignments and bi-weekly projects, we were also required to hand-in weekly “Outcomes” assignments, including Brand Statements, updated resumes, LinkedIn profiles, and written responses to questions like, “Tell me about yourself.”
A little Commencement Speech
Matriculating through this bootcamp was one of the harder things I’ve ever done. Unlike college where I had at most three one hour long lectures every day, in the bootcamp we consistently had about four hours of lecture followed by about two hours of review, in addition to another two hours or so of workshop time. In 12 weeks, we had to complete 20 out of 25 (80%) labs in addition to four projects and a capstone, which averages to about two assignments due per week. It was a pretty grueling pace made more difficult by the fact that these were topics I had never been exposed to before.
There are also personal and financial considerations. I don’t think I saw my friends more than once during the entire three months I was in the course because I spent my nights and weekends working on labs and/or projects. I hired a dog walker to help out with my big girl and ended up enrolling in MealPal because I just didn’t have time to cook. By Week 8, I was exhausted struggling to keep up with assignments, and looking down the barrel of a whole other month. In Week 9, my 10-year old computer broke. (Fortunately, I had my work backed up.)
With all of that being said though, if you are thinking of enrolling in a bootcamp, I would recommend it. While I am still learning and improving my Data Science skills everyday, I would absolutely say that the bootcamp gave me a foundation in programming that I previously thought unattainable. Before GA, I had never coded in my life. In 12 weeks, I was able to build predictive models, scrape the web, apply natural language processing, analyze complex datasets, and probably most importantly, gain confidence that I could learn programming techniques. My unsolicited advice to anyone considering one of these programs is:
Get help with the household (kids, dog, cleaning, meal prep) if you swing it/afford it. Every task taken off of your plate during this time will help relieve pressure.
Prepare financially to cover living expenses for the three months while in bootcamp and the three months after while job searching. (My experience thus far has been that jobs worth getting move slowly.) General Assembly (and maybe other bootcamps) offers loans for tuition as well as living expenses, if, like me, you don’t have about 25k cash lying around while living in NYC.
And finally, find your grit. Unless you are a data ninja before enrolling in the course (to which I would then wonder why you would feel the need for such a course to being with) it’s going to be hard. You will feel pressed for time, constantly. There will be concepts that you will not get right away. You will get stuck in your code. You may even fail a quiz, or five. This is a high-stakes environment given the cost of the class, the pace of the syllabus, the difficulty of the concepts, and the pressure to secure employment after. Knowing in your gut why you are doing this and what it will mean for you after - higher income, greater job stability, maybe even a better work/life balance - will carry you through. That, and a healthy supply of snacks.
If I can answer any more questions specifically about my own experience as a now alum of GA’s Data Science Immersive course, feel free to reach out to me via my contact form.
Here’s a picture of my cohort on our (very long) last day of class after 20 capstone presentations!
