You May Know Data Science, But You Don’t Think Like a Data Scientist
Discover the holy trinity of thinking like a data scientist
And now let’s discuss the importance of thinking and having the mindset of a data scientist, how you can do so, and the exact steps you need to take.
From thinking like an engineer, a scientist, and a businessman, to how these three different roles play an important part in becoming a data scientist, we’ll be getting your mind ready to face the challenges of a data scientist.
If you want to become a data scientist and create a project that will get you hired as a data scientist — you need to start thinking like a data scientist.
There’s a unique mindset you need to adopt.
And this mindset is what makes careers in data science so special.
To think like a data scientist you have to think like a scientist, think like an engineer, and think like a businessperson.
Think Like a Scientist
You’re going to be using advanced analytics, investigating motivations, looking for patterns, and trying to understand the WHY of it all in a systematic manner.
Scientists use critical thinking to discover knowledge using the scientific method.
So just like a scientist, you start by researching.
This means investigating the problem, finding out what’s already been done, speaking with stakeholders to see what they know, determining what’s already out there.
You’ll have to hypothesize and come up with some possible solutions or explanations for the problem you’re working on, thinking of ways to gain new knowledge.
You’re going to test the hypothesis, using statistical methodology to confirm or disprove the hypothesis.
And you’re going to measure your results very carefully.
Most importantly, you always have to keep an open mind.
Question everything.
Challenge and scrutinize your assumptions.
Suspend your beliefs and you focus on what the results are actually telling you.
Above all else, remember that thinking like a scientist is about being a critical thinker, not making assumptions, questioning and testing everything, and answering with data.
You must respect the facts.You must respect empirical evidence and accept the results of the process.
Think Like an Engineer
You have to write code, build systems, deploy models, and automate things.
You need to engineer systems that are reproducible, testable, extensible, portable, robust, reliable, and efficient.
You need to build systems to solve problems.
Avoid copying and pasting your work all the time.
You want systems that solve the problems for you.
You must remember that as a datas scientist you are writing statistical software. Avoid lengthy scripts as much as possible.
You must organize your code into well-documented software that can be automated if necessary.
This means creating pipelines, using object oriented design, and always with the goal of eventually scaling it.
By creating systems your code is easier to test and mistakes are more easily remedied.
You make your code scalable and extensible.
That way you can bring on new data sources and reuse your model pipelines for new clients in different industries, in different markets.
When you think like an engineer, you design solutions that solve the underlying problem.
This is what being a data scientist is about.
This is what thinking like an engineer is about.
This will make you much more powerful as a data scientist.
Use your project as an opportunity to build the skill of designing solutions that solve the underlying problem and aren’t just hacked together solutions.
No copy-pasta.
No just adding junk to scripts to fix solutions.
Refactor your code, make changes that solve the underly
Simplicity is better than complexity.
Build your projects on a solid foundation.
Avoid building something incredibly complex with numerous intricacies and many moving parts.
Think Like a Businessperson
Business is about making money.
That’s the goal of any business: make money, or spend less of it.
If you don’t make money, you go out of business, too.
Your job as a datas scientist is to solve problems.
And believe it or not — the problems you’re solving on Hacker Rank, or Leet Code, or whatever other site you’re using to prep for interviews are absolutely unrelated to the real world.
You need to demonstrate that you have business acumen.
That you can identify real world problems, and come up with real world solutions that solve them.
That’s it.
You don’t get a paycheck because you know 100 different algorithms, or because you know all of the ways to solve HackerRank problems, you get money to solve the problems.
Realize that as a business person, you’ve got to understand how to develop, quantify, measure, and track KPIs.
Business problems get complex and the way to understand if your business is headed in the right direction is by measuring key performance indicators.
KPIs provide the business with a way to monitor and gauge business health to gauge how successful it is.
Any particular business will have key performance indicators that they are tracking and measuring.
You must be aware of these as a data scientist.
What are those KPIs for this business?
What does this business care about?
How do those predict or measure this business’s success?
You must also appreciate that business is about relationships, and the best way to build relationships is through clear and empathetic communication.
Make sure you’re communicating in an easy to understand language and that will make your audience feel like they understand what you’re doing — to the point that they may even think to themselves: “Hey, I get this. I can be a data scientist!”.
And this is the trifecta of the data science mindset.
That’s it for this rant. I’ll see you all in the next one.
Let me know what you think. Leave a comment below, let’s open this up for conversation.
I‘ve also got a free, open Slack community where I’m happy to bounce around project ideas with you!
And remember my friends: You’ve got one life on this planet, why not try to do something big?
