Should you learn data science?

Nearly everything you do generates data. Visit a website: Data. Tap an app on your phone: Data. Buy something with a credit card: Data. Billions of people are generating immense amounts of data every single moment of every single day. That’s some big data, and it’s only getting bigger.

What exactly is data science?

Now, you might have trillions of rows upon rows of data showing what customers are buying or how pollution around the world is changing, but none of that is useful on its own. In fact, it’s overwhelming. Go ahead and try to read through a dataset with billions of rows in it. We’ll wait…. Ok, please come back. The point is that this data, on its own, is unusable. It takes work and specialized skills to transform it from unintelligible noise into something that can be easily understood.

Enter data science. Data science is all about diving into that well of information and shaping it into a tool that you can use to accomplish a goal. Data scientists can process data so it’s human-readable, building visualizations that tell a story. Sometimes we use models that explain a process or predict a behavior. Other times we run experiments to validate the things we think might be true but want to take a step towards actually being able to prove. The essence is that we intake raw data and output something that is valuable: you can do something or learn something from it.

What skills do you need as a Data scientist?

Does this sound exciting? It does to us. Data science is a rapidly expanding field that’s tackling some of the biggest problems in the world today. But what does it take to actually be a data scientist?

Your most important skill is to find solvable problems. Data science is rarely cut and dry. It isn’t simply "apply this technique" or "run this program". That’s necessary, yes, but that's usually the easy part. You need a thorough understanding of the problem so that you can determine which tools are best suited to your task. Learning data science, then, is not merely combining programming with statistics — it includes that, but also requires context. You need to understand the domain that you’re working in, so you can test your hypotheses in the real world.

Before we go too far, let’s talk about that core toolset. What specific skills do you need to know? Certainly you need some analytic techniques: how to make visualizations, use summary statistics, and the like. You probably need to be able to run tests, so understand things like A/B testing and statistical significance. That way you can start to prove if the things you’re doing, either as a company or an individual or whatever you’re working on behalf of, are creating the responses they’re designed to create. You'll also need modeling and predictive skills for a lot of roles. Certain problems will also require a deep statistical and mathematical knowledge or serious programming skills, particularly if you have to develop your own algorithms or apply them to a more rigid statistical framework.

Beyond that, what you really need is a passion for finding and solving problems. You have to want to dive into a space and stay a while, figuring out its eccentricities and peculiarities. That kind of curiosity is the most valuable resource a data scientist can possess.

How can I learn data science?

So maybe this is starting to sound exciting. You’ve read this far after all. But how can you actually become a data scientist? Learning anything requires a positive feedback loop. Find an appropriate way for you, based on support, structure, and opportunity to apply the skills you’re learning.

In designing our bootcamp at Thinkful, we've found that students learn best with:

We've found this combination to be ideal in learning critical data science skills.

Working with Thinkful, you’ll build a set of skills that allow you to approach a series of problems in a field you care about, building reports or models or product that takes data and makes life better. If you are interested in learning more, take a look at our Data Science bootcamp.