DJ Patil and Jeff Hammerbacher coined the title Data Scientist while working at LinkedIn and Facebook, respectively, to mean someone who “uses data to interact with the world, study it and try to come up with new things.”
Although they purposefully kept the term vague to encapsulate the broad nature of the field, a common argument goes:
“Data science is a science, so you need to be an experienced PhD to be a practioner.”
Others simply denounce “fake data scientists” with “no knowledge of the fundamental theory of the field.” As another article puts it, “data scientists must learn Python.”
However, we can clearly see in a few ways that a data scientist doesn’t need an advanced degree, coding expertise, or even a lot of experience:
Remember how analysts at Facebook and LinkedIn coined the term “data scientist?” Well, PhDs (or even Master’s) aren’t a requirement at either of those companies, or almost any of the 100+ companies hiring data scientists right now.
As a Facebook exec writes:
“We’re looking for talent, no matter what their background.”
Indeed, Zuckerberg himself is a drop-out.
If we look at a Data Scientist job opening at Facebook, we’ll find BS “or equivalent experience” in the job requirements.
The same is true for this “Product Analytics Lead, Data Science” role at Google — asking for a “BA/BS degree or equivalent practical experience.”
The “or equivalent practical experience” is featured after the preferred degree in every opening I looked at, even in a Data Scientist Lead role.
Looking at similar roles at LinkedIn, there are asks for a “BS or equivalent experience.”
Like Google, Netflix doesn’t require a degree at all.
In the Internet’s early days, building a website required serious technical prowess. Now, no-code tools like Wordpress and Wix enable anyone to quickly launch a site.
This “democratization” process has occurred in many fields, including design (Canva), information (Google), computing (Apple), commerce (eBay, Shopify), and now in data science, with no-code tools like Apteo.
While the early days of data science required considerable experience across a range of fields, today’s solutions are more “plug-and-play” than ever.
Now, anyone can “turn data into insights,” including making visualizations and predictions on data, without any code.
By now, you might be saying:
“Sure, companies don’t need advanced degrees, but those with advanced degrees make better data scientists.”
This is also wrong.
Data Scientists like Nate Silver and Paul DePodesta, who helped popularize the field, have bachelor’s degrees in economics, but no advanced degrees. Hammerbacher, who we mentioned earlier as he helped coin the term “data scientist,” only has a bachelor’s in mathematics.
Kaggle’s top competitor doesn’t have a Ph.D., and Kaggle’s CEO himself admonishes the lack of practical thinking he sees in some data science PhDs:
“Ph.D.s in computer science and statistics spend too much time thinking about what algorithm to apply and not enough thinking about common sense issues like which set of variables (or features) are most likely to be important.”
There are many bootcamps or “skills training” courses that teach the skills necessary to be a data scientist, regardless of their academic background.
In fact, these are extremely valuable for the industry, as they help to fill the shortage of tens of thousands of analytics professionals.
For instance, BrainStation has taught 75,000+ students, with graduates at companies like Facebook, Microsoft, IBM, Spotify, and Amazon.
FlatIron’s data science program teaches the skills you need in 15 weeks, and even guarantees your money back if you don’t get a data science job. So much for “you need an eight-year Ph.D.”
Other data science bootcamps, like Galvanize, boast that over 2,250 of their graduates have been hired.
Although some point to a Ph.D. as a “holy grail” to enter the field, this couldn’t be further from the truth.
Even with a Ph.D., you’d still need to acquire practical skills to work in the field. As one Ph.D. graduate wrote:
“I had just walked away from 8 years of study and hard work with no plan.”
In 6 months, she taught herself the practical skills to become a data scientist, as any Ph.D. would need to do.
In other words: “A Ph.D. is not enough,” and many books and courses exist to re-train academicians to acquire more practical skills.
Many PhD data scientists will admit that “you don’t need a PhD to do data science,” and there are several factors that can hold back PhDs, such as:
You absolutely do not need a Ph.D. to be a data scientist, and while it can help, it can also hold you back.