My takeaways from winning many data science competitions (with grand prize!)
In 2018, I was admitted to the Master of Business Analytics program at MIT. The admission notice first came as a phone call, starting with the area code 617, from Cambridge Massachusetts. Many years later, I still cannot describe the excitement at that moment.
That was 4 am in Shanghai.
I was still awake,
anxiously waiting for the phone call.
At that time, it was beyond my wildest imagination, since the program notoriously had never admitted anyone who accomplished undergraduate study in mainland China, in its first 3 years. I was the first one. In hindsight, I was the only one in the program’s first 5 years of admission.
In retrospect, among the many reasons that contributed to my admission, I think that “X” factor, the one strength on my resume that many other candidates wouldn’t have presented, might have been my endless participations in data science competitions, which manifested my great passion for data science — not to mention that I also luckily won some of them.
I didn’t keep the nice photos of all the competitions I have attended. This article only includes my most loved ones:
Wining Citadel’s Data Open at MIT with a $20,000 USD prize
Wining 2nd place among 100+ teams (from best universities in China) in an Algo trading competition in Hong Kong along with $35,000 HKD
Wining 3rd place in the National SAS data mining competition in Beijing out of 1000+ teams
The benefits of attending data science competitions for a future-data scientist
The benefits of attending these data science competitions are multifold. I sincerely encourage anyone who wants to become a data scientist to start doing it.
First, you had the opportunity to learn new skills. In most of the data science competitions, I had to acquire new skills or learn new algorithms in order to achieve the best results. The competitions are usually mentally challenging,
Second, in most of these competitions, you will have the chance to work in a team setting, surrounded by talented people, under an urgent deadline. This perfectly simulates your future work environment (where you work in a team setting, surrounded by talented people, to meet a not-neccessarily-urgent-but -more-important delivery date). You will gradually learn teamwork, time allocation, some-what project management, and effective communication. All these,
One last benefit, which also comes a bonus point that I discovered, is that these competitions became great channels to get to know more people that share the same interests as me. I met with a nice lady from SAS company in Beijing during my competition, and she ended up writing an inofficial recommendation email for me during my graduate school application. One of the closest mentors I have right now at Amazon (who is a data science manager), was introduced by a friend I made at one of those competitions.
Oh, and of course, these competitions sometimes offer grand prizes. For me, that helped a lot throughout my poor student life. For example, my college tuition was around 5000 RMB (which roughly equals 5500 HKD). The 35000 HKD prize to be split among 4 teammates might seem not worth mentioning (if you calculate into USD), but easily afforded my college tuition for more than 1 year.
Competing in Citadel’s DataOpen world final, did not win a penny. However, it was an interesting experience. We published a paper later on with our invention during this competition
Winning 2nd place among 30 teams, not a competition, but we made a huge impact helping our client Schneider Electric
Takeaways from my competition experience
Back to my title, what are my learnings from these competitions?
The first and foremost is teamwork. I always meet with my teammates before the competition to discuss work assignments and the collaboration pattern. For example, in the Algo trading competition, I worked on integrating the strategy with the trading platform and made the final presentation. My teammates handle the strategy and I fully trusted them. In my SAS competition, I took the team leader role and helped to organize the work. These made our collaboration seamless and optimized the time usage. The most amazing part of these competitions was seeing the heated discussion where endless new ideas were generated, and the smooth execution that gave me a sense of belonging to the team.
Second, a lot of these competitions that I participated come with a presentation (whether in the form of a report or an oral presentation). Writing a good paper or giving an impressive presentation is always more important than having a 0.1% performance lift. Think about the rationales behind your analysis and how to convince people. These are the real-life scenarios you most likely will be dealing with when you start
Third, in those competitions with a time constraint, really get familiar with your tools and skills. Working under a time constraint, there is no time to search google for basic syntax. I personally maintain a library of my frequently used algorithms and visualization code chunks, which equips me with the capability to conduct fast data analysis in either competitions or normal work settings.
In those competitions with abundant time, don’t be afraid to try new ideas. It’s a great opportunity to learn. That way, even if you are not earning a penny, you can still work off satisfactorily with your new learnings and new findings.
Thank you for reading my article until this line.
My name is Andrew Huang (aka. Drew). I graduated from MIT. Currently, I am an Applied Scientist II at Amazon. I wish to grow into a chief data scientist one day.
I haven’t been updating my personal blog for a while. The trigger of me writing this document was simply because, I just earned another #1 in a recent Amazon internal data science competition, which brings back my old memories and incents me to write this article and share my personal experiences and suggestions with more data science enthusiasts.
I have written a lot in Chinese, having 40k subscribers on China’s medium counterpart. I will write more articles in English, and I will start to do more Kaggle competitions. Look forward to sharing my new Kaggle experiences with you in the future.
Please consider subscribing.
