Hi everyone! My name is Michael, my friends often call me Mike or MT. And of course, I’m just another aspiring Data Scientist just like the most of you out there on this platform. But unlike some of you, I was also an ex-biomedical engineer. I am interested in all things data-driven, but particularly in uncovering patterns that have temporal attributes, such as trends and sequences. And because of my prior educational and industry experiences in the healthcare, I’m really passionate about constructing scalable data solutions that are also socially meaningful.
Just a bit about my Medium
At the time of this post, I have been a Medium subscriber for over a year. I started because I saw an incredibly diverse community where people like me share their experiences and knowledge, which I found really inspiring and motivational. And I think now is a perfect time for me to start my own blogs here for the reasons I will describe shortly. However, unlike most educational Medium articles that cover latest breakthroughs in AI and state of the art algorithms, I’m writing mainly to review concepts in stats, ML and topics that I personally found interesting or confusing. In other words, I don’t expect many readers unless you have the magical link to my page. The point is, my blogs might not be the most comprehensive and visually pleasing, simply because I’d like to keep the formatting simple and the articles short most of the time. But I will occasionally share some projects that I’m working on.
And a bit more about my background
If you happen to be reading this article and wonder what I’m up to right now, I’m currently a 1st year graduate student pursuing a degree in Data Science at Duke University. Like I mentioned earlier, I already have a Master Degree in Biomedical Engineering from the University of Toronto. And I also spent over 1 year as a Product Development Engineer at a medical device startup company back in Toronto, where I originally came from. “MT, that sounds like a promising career, why switch to Data Science and have you not had enough school yet?” That’s the reaction I got from most of my friends. Well, it wasn’t an impulsive decision, and I will explain that in the rest of this post series.
What is Data Science to me
Data Science has become a buzz word among many college students and professionals for the past couple of years. Many hope to become a Data Scientist for its lucrative compensation, and that was also my impression when I first became aware of this profession in my last year of undergraduate study back in 2017. However, I chose to ignore it because I thought it was simply another fancy derivative of a Computer Scientist, plus I was determined to get a graduate degree in biomedical engineering. That plan worked out pretty well, but little did I know that it was also where my journey in Data Science began.
My first true encounter with Data Science was from a side hustle after I joined the workforce, which was an experience that really made me appreciate the power of data-driven technologies. I was looking for some projects that could keep myself occupied during my spare time and could also offer decent financial returns (I probably cared more about the latter at the time). That’s when I found out about drop-shipping. If you have never heard of this term before, it’s basically an e-commerce strategy where you curate merchandises at low prices and have the manufacturers ship the products directly to the customers. Drop-shipping is highly favoured by small business owners, because of the minimal upfront investment since there is no need to stock products. And there are no distributors and retailers in the middle, which means you could keep all the profits. All you really need is a laptop with stable internet access. Sounds like a get-rich-quickly scam? It depends. But the point here is that it introduced me to digital online marketing, which is just big data combined with machine learning in its core. Specifically, to get my products some public exposure, I had to run advertising campaigns on Facebook and Google, and utilize various kinds of data collected from tools such as Facebook Pixel to determine the user demographics I should target my ads at. And of course, the more ‘accurate’ my estimated audiences, the more likely my products would sell, and the more I could make! This was mind-blowing to me and completely transformed my understanding about data science.
Meanwhile, my company was developing MRI-compatible infotainment systems. Oftentimes, patients with anxiety or claustrophobia struggle with MRI scans due to tight space inside a bore as well as extremely loud noises produced by the scanner. Our system attempted to alleviate these issues by distracting patients with media contents sourced from providers such as YouTube and Netflix. This might sound gimmicky, but it was a really big deal for a lot of patients. Also, making electronics work inside and across a faraday cage was not easy, let alone a magnetic field that could go up to 3T. So intuitively, this system was perfect for neural marketing research — researchers could use our system to present a collection of products, advertising or marketing elements to a human subject during a fMRI session, and observe this person’s brain activity in response to these catalogues in real-time. The goal of studying these physiological changes is to predict consumer behaviour and bolster decision making (e.g. set pricing and improve branding). This process requires sophisticated data analytics and modelling techniques that I was unfamiliar with at the time (still am), but the technology and the idea behind it were eye-opening to me.
At the point in time, Data Science seemed powerful but manipulative. It also felt ubiquitous yet inaccessible at the same time. Its mysteriousness really appealed to me and pushed me to further investigate the technology behind it.
It seems like this post is starting to get too long, so I will continue my story another day.
Stay tuned! (if you read)