I took an intensive 12-week data science course. Here’s what the experience was like.
This article was originally authored by Anirudh Kashyap and published by General Assembly here.
As I approach the end of my Data Science Immersive (DSI) course at General Assembly in Washington, D.C., I’m reflecting on the past 12 weeks, which have been a whirlwind adventure of hard work and perseverance. After three months of intense training, it’s rewarding to sit back and evaluate the journey of going through a data science career accelerator, also referred to as a bootcamp, and what have I gained from the course.
As a 30-year-old going back into school, I was apprehensive and uncertain taking this course. It was out of my comfort zone, not in my previous field of study (engineering), and I was unsure of its effectiveness.
However, it was worth it. Diving into a full-time data science program has been a really positive experience for me. While a full-time course might not be for everyone, my hope is that sharing my experiences can help others who are considering an education in data to get ahead in their career.
1. A full-time course is a serious full-time commitment. Get ready to learn a lot, very quickly.
During the orientation, my instructors at General Assembly described the course as “drinking out of a firehose.” It’s a really intensive program that hits you with a ton of information and training, and you have to be ready to take it all in. For 12 weeks, your life, TV, household chores, friends, and gym take a backseat. If you’re in a relationship, it’s a commitment that ideally should be made with your partner, who will need to understand how rigorous this undertaking is.
Once you’ve made the decision to partake in this journey, you’ll very quickly realize that you need to do a lot of work — and fast. There are a whole lot of labs, projects, blog posts, and quizzes to tackle on a weekly basis.
Asking questions and reviewing course materials daily is important in order to reach the destination without hiccups. Monday Night Football can wait — reviewing course materials is top priority every evening.
At some point in the program, you may hit a wall — and that is OK. You may realize that some topics are not made for you — and that is OK. You may realize that you don’t remember TensorFlow after learning it a week ago — and that is OK, too. You will be great at some modeling techniques and be able to explain them to a layperson in a 500-word essay — and other techniques you will read about and most likely forget in a few weeks. Also OK. You need to pick your battles.
This is all part of the process as you discover your strengths and interests, and flex new data muscles in your brain. If artificial intelligence or image processing are paths you want to explore in your career, then understanding and getting comfortable with data science technology like neural networks and TensorFlow is paramount. But, the majority of data scientists in the industry are not building AI-assisted cars — they’re leveraging data to advance in a broad range of interesting industries. A survey from the data science platform Kaggle shows that (believe it or not!) skills like logistic regression and decision trees are many data scientists’ bread and butter. Throughout DSI, we had a ton of guest speakers who mentioned that exploratory data analysis (EDA) makes up a ton of their efforts on the job.
The main takeaway? You’ll gain comprehensive knowledge about the topics you are comfortable with and consider to be your strengths. On the road to getting there, you’ll also be introduced to plenty of tools, techniques, and strategies that will likely pop up — and be helpful — in a lot of different jobs.
2. You’ll meet a ton of great new people. Learn from their diverse experiences and use them to build your network.
An immersive education program has students from different backgrounds, and most of them are career-switchers. Each person is a great resource of information, and I had the opportunity to learn from fellow students and hear their varied perspectives. Watching my classmates, who have experience in facilities, journalism, actuary, and economics, tackle a problem from different angles was a great learning experience. I have improved in many aspects of project management apart from model building — including presenting, problem-solving, and scheduling — by learning from my peers.
You’ll also automatically expand your LinkedIn and in-person network by being a part of these programs.
Meeting people from different background offers connections in fields much different from your own. I come from an engineering background and now I have connections in many other industries who will be focused on making moves in the data science field. You’ll see a funnel effect where people from various sectors study together, and then bring their new data skills to a range of roles in other fields. This will be your new network, and tons of opportunities may come from your connections.
When I look at data science roles on LinkedIn, I’m connected to more people in a particular company through General Assembly than through my graduate school, University of California, Irvine.
3. Your instructors can make or break the experience.
This one is a cliche and can be said of any educational program. Just like a bad boss can make you hate your job, a bad instructor can break this experience for you. Fortunately for me, my instructors have been excellent and were able to distill difficult concepts like Bayes statistics into simple, lucid form. Smart folks are not always great teachers, but my instructors at General Assembly were both.
A great instructor should have subject-matter expertise and the ability to distill it down for folks new to the topics. They should repeat themselves, and if students are confused, try other techniques to make sure they understand the content. They also need to have patience, and make sure students never feel stupid asking questions. All questions should be answered so that everyone in the class understands.
My instructor, Matt Brems, broke down Bayes equation terms like posterior probability, marginal likelihood, and prior probability, repeating them every day for the first week. Every single day, in both the classroom and during office hours. In the end, I can close my eyes and explain core statistics concepts to people in ELI5 (“explain it like I’m 5”) technique.
Please scout your instructors and know more about them before joining class. I was lucky, but make sure you do your research!
4. Choose your final project wisely.
Like many long-form classes, DSI requires students to complete a capstone project that demonstrates the skills learned throughout the course, showcases the branches of data science you want to pursue in your career, and is a culmination of 12 weeks of coursework.
My friend Ben, in his project, wrote an excellent blog post about his approach to solving a data science problem, and explained the rationale behind selecting a capstone project with practical implications. His project detailed how he uses data to help his moving company make better estimates for the cost of moving from point A to point B in Washington, D.C. Similarly, my project’s purpose was to help a business, The Whisky Exchange, understand how a company should price its whisky.
The capstone project at the end of the course should accomplish three goals:
- It should reinforce your data science toolkit. I strongly believe that your capstone project should encompass the data science tools that interest you and the tools you want to pursue in your career as a data scientist.
- It should be relevant to a business problem/group that needs some solutions.
- It should be fun! After all, data science should be fun and learning cool things using data is a very satisfying experience.
If it achieves all of the above, you’ll have a great portfolio piece that showcases your interests, skills, and talents. This will be a great resource to share with future employers, collaborators, and data nerds who want to get to know you better.
My capstone project was about understanding the factors that affect the price of whisky. I collect whisky and my project was done in collaboration with The Whisky Exchange to understand what sets expensive whisky apart. It was a lot of fun and I was able to use natural language processing, ensemble methods, and logistic regression to understand the labels, flavor influence, and other factors that you find in cheap or expensive whisky.
It was fun, and I learned a lot.
Some parting words: GA’s Data Science Immersive has opened doors for me that would not have been possible with my graduate degree in chemical engineering. It has helped me make connections and change my career. I would highly recommend it if you have curiosity into the mysterious field of data science.
Dive into full-time learning at General Assembly. Learn the skills to become a data scientist, full-stack web developer, or user experience designer in our full-time, career-transforming Immersive programs. If you’re looking to leverage the power of data in your current career, explore our part-time courses in Data Science and Data Analytics, or learn online with our Data Analysis course.