Becoming a Data Science unicorn – Part 1

At the beginning of my data science path, three years ago (decades in data science years) the most frequent question I kept getting asked was “What on earth is data science?”.

A lot has changed since then. Now, I believe, almost everyone has heard the terms data science and Artificial Intelligence. What hasn’t changed is that I am still getting asked “what is data science”, a lot!

My answer has evolved a lot with every iteration, from a detailed definition of what machine learning, statistics and data analysis are to the answer I give most often these days: “Data science is the discipline of making data useful.”

Although this definition may seem quite vague, it is an umbrella term, which means something different for every company and is used to describe pretty much everything, from data and business analytics to data engineering, big data, machine learning, deep learning…

Hype around all of these terms resulted in a new set of questions that I get asked a lot these days – “Should I pursue a career in data science?” and “Where do I start with data science?”.

If I cannot give you an exact answer to the first question, I most definitely cannot give you a definitive answer to the latter ones. However, as someone who asked herself these same questions numerous times (even after starting my data science path), I understand the importance and burden of such questions.

So, even though I cannot offer you enlightening answers, what I will try to do through this blog series is help you find your own answers and support you on your data science journey. I will do this by sharing experiences we at Atlantbh collected while working on our very own data science project PlaceLab, as well as those gained during numerous data science internships  and by revealing lessons I personally learned the hard way.

Should I start with data science?

Honestly, I do not know if starting with data science was harder at the beginning of my journey, when it was really difficult to find relevant resources, or today when you are bombarded with numerous papers, must read materials, must do courses, must know terms, frameworks, tools, must have skills…

Before tackling these issues, there is one more stone to overturn – to decide whether you should pursue the data science path or not.

Two common dilemmas exist when contemplating this decision. The first dilemma is knowing if data science is a smart move or not; is it a viable career path or just a bubble about to burst. The second one is determining whether or not you have what it takes to become a data scientist.

I would not argue the first point too much, I think the latest AI achievements in the industry, healthcare and everyday life speak for themselves. The fact that all of us carry numerous AI algorithms in our pockets is proof enough that we are already living in the era of AI. My personal opinion is that this is the best possible time to start with data science – it’s the beginning of this exciting era, but also a time where we know, with enough confidence, that AI is here to stay.

The second dilemma is usually the one which causes people to give up on pursuing a data science career. Definitions for data scientist found online usually present some magical creature with a google-level of software engineering skills, Jerome H. Friedan level of statistics skills, Steve Jobs presentation skills and CEO business skills – in every domain. This image of a data scientist almost prevented me from applying for my current job. In hindsight, I am so glad that I never gave up and today I can tell you that at the beginning of your path, you can forget about this highly irrelevant image and honestly ask yourself if this is something you really want to do. Then ask yourself are you enthusiastic and excited about everything happening in the data science field? Are you intrigued with data? Do you love numbers, patterns and coding? Are you willing to work hard and improve constantly? And, most importantly, are you curious and a problem solver? If the answer to most of these questions is yes, then data science is definitely for you. However, even if you don’t resonate with most of these, the most important part is whether you really want it, if yes, then you should definitely try. Try it without letting anything, including your current skills and your background, stand in your way. Of course, due to your current level of skills, you might struggle a bit more with some things, but I believe that most skills needed in data science can be learned if one truly wants. Also, as Rachel nicely stated here,  if your background is not in the maths or computer science, you are exactly what the AI world needs because it needs more diversity in order to get to the right place.

Do not start just because

Having said that, even though I strongly believe that you can succeed in the world of data science if you really want to, I believe the opposite applies as well. So, do not start just because; just because it is a buzzword or because it seems fun. Don’t get me wrong – data science can be really fun, but you need a stronger WHY than that. You need a WHY strong enough to carry you through all the tough parts and obstacles because learning and practicing data science can be hard, like really hard. Not because of the math or aligning tensor dimensions – those are the easy parts. The hard part is staying motivated, dedicated and determined even when you get stuck while learning or on the project you are developing – and you will get stuck from time to time. Enough motivation to wake up earlier in the morning to keep up with the latest papers and resources, enough motivation to allow data science to become part of your daily life, enough motivation for lifelong learning to cover multidisciplinary skillsets and to keep up with fast-paced changes. Even when you do all of that, taking into account that as a (wannabe) data scientist you are probably someone who is at least to some extent a perfectionist and pretty hard on yourself, you will still experience imposter syndrome on steroids most of the days.

Also, do not start with the wrong expectations on how your data science job will look like. The truth is that you will probably not build flashy models all day long and change the world with every project you do. The steps necessary to create working projects include lots of parts before and after training a model and most of them are not sexy at all, despite what they keep telling us.

Wrapping up

For those of you who are still with me, a strong WHY will get you through all of this and at the end of the day, when you read about an exciting new model and realize you are part of THAT world or when your own model gives you the results you were aiming for, you will realize that data science is fun, stimulating, rewarding and worth it!

So, until next time, when we will try to cover how to actually start with data science, go in pursuit of finding your own WHY. Also, maybe take a peek at the great new course “AI for Everyone” from Andrew Ng.

Click here for Part 2.