How to Become a Data Scientist

Data scientist is the buzzword job title that automatically instills a level of coolness and expertise in anyone working in the tech field.

Because, let’s face it, explaining how companies can reach their goals with numbers is pretty cool.

Much of the academic and EdTech world agrees with the popularity of data science. With universities offering several advanced degrees and courses in data science and analysis, and more data science online classes popping up daily, it’s safe to say that the “currently unrecognized” science of data, as John Tukey predicted already predicted 50+ years ago, is here.

The US Bureau of Labor Statistics estimates jobs for computer and information research scientists (aka, data scientists) to jump 16% from 31,700 jobs (in 2018) to 37,000 come 2028, a higher-than-average growth rate.

But how do you get into this industry? What does the data scientist career path look like? What are the different gigs of data scientist that you can do? And what does it take to explain big-picture figures to company higher-ups and stakeholders? 

Here, I’ll show you some steps to help get you there, plus experienced pros weigh in with their career tips; that and stats, research, and more along the way!  

Know What Data Science Means

Data science means different things to different companies says Neal Lathia, Machine Learning Lead at Monzo.

Data science is like this big, umbrella term used for someone who digs through the numbers figuring out what they mean for the company and what next steps they should take. That “random email” about Product X from Company Y you’re running out of likely found its way into your inbox thanks to data scientists.

But data scientists do more than just marketing; a team of academics, for instance, sifted through 2.3 million Google Scholar papers to find out how often the search engine directs readers to free research papers. (They found 55% of all docs published 2009 and 2014 were free to read via Google Scholar.) As one Redditor points out, a data scientist is a “solutions developer.” 

Depending on factors like specific company and job description, one data scientist may differ from the next: “Depending on who they are and what they need—what you could do in your day-to-day as a data scientist could range from analytics to stats or machine learning’” Lathia says. He recommends that you always ask companies what data science means to them when you apply.

Also, one person working as a data scientist in one company may have the same job responsibilities as someone who has a different job title. (That US Bureau job openings prediction we mentioned earlier could be on the conservative side.)

There’s been a debate in statistics circles, for instance, whether data scientists are really statisticians with a different job title. Another topic for another day—Basically, if you like solving big-picture problems with numbers, you’re in the right place. 

Get That Bachelors, At Least; Maybe a Masters Too

One study looking at 1,001 data scientists’ LinkedIn profiles revealed almost three-quarters hold a Masters (74%) and a little over one-quarter have a Ph.D. (28%). Nineteen-percent had a bachelors’. 

The US Bureau of Labor Statistics supports this, citing a master’s degree as the norm. We did a rough job search ourselves on Indeed for data scientists and found most companies require an advanced degree. Not all, but most. 

So, bachelors, yes. Masters, most likely. Ph.D., could be a good idea. 

As for the specific degree, the study (surprisingly?) found it was across the board: out of the 1,000+ data scientists,

Data Scientist, Carla Genry recommends a mathematics base, “[I]f you have a mathematics background, you’ll always be ahead of the crowd!”

This could be a traditional quantitative degree or self-taught courses and experience.

Meet Your Best Friends, Python & R

“The data scientist is only as strong as their toolset. Python skills are, by far, one of the strongest tools a DS can have in their toolbox today,” one professional data scientist says. 

Python is a programming language that most data scientists know. Hard science or soft science degree, learning Python is key, R too, the second most popular. Some data scientists may have learned these in school others are self-taught. 

The same Redditor mentioned he self-learned Python and SQL from CodeAcademy and YouTube. It’s definitely doable and possible, and there are online communities like GitHub to practice and seek mentorship from.

Other skills to have, in addition to Python (the most popular), as one poll lists include:

  • Data visualization
  • Critical thinking
  • Communication skills

Soft Skills Are Important

One skill set we kept hearing from experts and seeing on company job postings was communication. Being able to communicate your findings in laymen’s terms to hire-ups and stakeholders are data scientists’ bread-and-butter.

Because this is what moves mountains: “A lot of online courses focus on machine learning. However, the most impactful data scientists I’ve come across are the ones who can impact business and product decisions using insights from data,” Lathia says. 

“For folks starting out on their careers, I usually encourage them to think more about impact than the shiniest new tools in machine learning.” 

Another data scientist echoed a similar sentiment, “A data scientist should strive to become the bridge between the business community and technology community.” 

Taking the Road Traveled and Less Traveled By: Both Work in Becoming a Data Scientist

While data science is new and comes across as hip and cool, Gentry, says that’s not what it is at its core, “It’s still about finding trends and information that can help solve a problem.” That’s the really cool stuff. 

You can get there via a combination of education and internships, some job experience and online courses. Remember, it’s all about the numbers telling the big picture. 

Questions? Comment below!

Commentary are personal opinions and does not reflect the hiring practices/policies of organizations mentioned.

Join the discussion

Menu