4 Things I’ve Learned During My First 8 Months as a Data Scientist (In Crappy Drawings)
I have learned so much in the past 8 months, and I was initially going to write another “What I Learned From Being A Data Scientist for <insert length of time here>” article full of fancy words and paragraphs that people have seen a gazillion times. Clearly this has been done before, so why not try to do something new? Something more creative?
Oddly enough, I’ve built a reputation at work for using hand-drawn diagrams in all of my presentations, from weekly touch-points with our lead data scientist to technical design meetings with audiences of over 100 people, including engineers, product managers, and other directors of fancy things . I figured if I can incorporate my crappy drawings into company-wide presentations, I might as well incorporate them in my Medium articles too!
So here we go, the four biggest lessons I have learned as a data scientist (portrayed via my embarrassingly bad drawing skills).
Please note this article is based on my personal experience, so feel free to completely disregard everything I say from this point on.
Your Role Is Related To Your Company’s “Data Maturity”
The x-axis represents the possible range of a company’s “Data Maturity”. If your company knows it has data but doesn’t know what to do with it, then your company is in its data-adolescence, and it’s going to need you, the data scientist, to do A LOT MORE than run a few regressions. On the other hand, if your company has been around the data-science-block a couple of times, you might not be called on to bring a machine learning model up to scale or design the next AI-enabled product. But you’ll probably be required to conduct business-value-driven studies that will help bring your organization closer to its goal(s).
When All Else Fails, Throw Some Spaghetti
I’ll never forget one of the first conversations I had with my manager:
Manager: “Do you know what you’re here to do?”
Me: “Uhhh yes. I’m here to create the company’s anomaly detection system…right?”
Manager: “Right. Do you know how you’re going to do that?”
Me: “To be honest, I don’t” (Honesty is the best policy)
Manager: “Great. I don’t either. So we’re just going to throw spaghetti on the wall and see what sticks.”
I get the sense some people feel like data scientists know all the answers to everything. I know I don’t. Sometimes the only thing you can do is throw all of your proverbial data science solutions spaghetti at the proverbial problem-wall and see what sticks. If nothing sticks, then you go do some more research, gather up another plate-full of spaghetti, and make another mess :)
It Can Be Extremely Overwhelming
Slightly related to what I said previously, I know I have put a lot of pressure on myself in the past to know everything there is to know about data science. All the algorithms, the stats, the math. The list goes on and on. The truth of the matter is it’s simply impossible to keep up with everything that’s going on in the data science world. Numerous papers on the latest and greatest algorithm/model are published every day. New technologies seem to pop up every month. It’s overwhelming dude! The good news is once you realize it’s impossible to know everything, you can change your mindset. The way I see it, as data scientists, it is not our jobs to know everything, but it is our job to be able to use everything. We need to be able learn unfamiliar techniques and methodologies to solve novel problems. We need to be able to use all the tools available, even if we don’t initially know exactly what they are. We welcome failure but demand success. And that, ladies and gentlemen, requires lots of spaghetti.
You Need To Be Good At Sculpting
As a data scientist, your responsibilities go far beyond the reaches of your computer. Not only is it your responsibility to solve problems, but you also need to present your findings to other people who, most likely, are not fellow data scientists. You need to be able to mold your findings, your ‘clay’, into shapes that fit into your audience’s realm of understanding. One day you’ll have to mold your clay into a triangle for management to firmly grasp the results of your latest study, and the next day you’ll rework your triangle into a circle so the engineering team can figure out how to put your results into action. Communication is key. The ability to mold your fancy data science ‘clay’ into shapes other people can understand and relate to is an extremely important skill for any data scientist to have. And personally, this is where I find the most joy in my role as a data scientist.
There’s only so much a scatter plot, two circles, some spaghetti, and a clay metaphor can convey about the lessons to be learned when navigating the crazy world of data science, but I’d like to think the messages they communicate are extremely powerful and humbling, though the quality of the drawings is abysmal.
What lessons have you learned since starting your data science journey? More importantly, how would you draw them out? Feel free to leave them in the comments below. I’d love to see them!
