It wasn’t too long ago that the job title of data scientist was invented. In the time since then, it’d be optimistic to say that we have a unified concept of what being a data scientist entails.
There are many valid questions worth asking if you want to arrive at an ontologically satisfying conclusion: What is data? What is science? Does data science become less scientific the deeper it falls into the corporate status quo?
A co-writer of mine at CirroLytix has put his stance forward concerning the nature of data science, but I won’t be weighing in on the same terms—not for now, anyway.
Instead, I’ll be looking into a practical question: In terms of job performance, what sets a good data scientist apart from a pretender?
Grab a notepad and a mirror, folks, because we’re about to get introspective.
You Fixate on the Practical
Elegant models and cutting edge AI deployment are well and good. The world needs people who go the extra mile to chase perfection and experiment with emerging technology.
The average business, however, doesn’t.
A data scientist worth hiring puts a premium on efficiency. They know better than to condescend when presented with the occasional redundancy, and they definitely value function over form.
This isn’t to say they’re lazy or prone to cutting corners. In fact, the hallmark of a good data scientist is their ability to generate and communicate clean and usable insights, not pretty charts.
Practicality can also be demonstrated by the tools you pick. Specifically, if you’ve ever been nitpicky or defensive about your tools, odds are they aren’t as qualified as you claim to be. Like any professional worth the name, a data scientist worth hiring would stick to tools which they’ve mastered, and which prove comfortable for them as well as the clients they’re servicing.
The closer you look to an arrogant nerd with a chip on your shoulder, the farther you are from being an asset to a project or company.
You Never Stop Learning
There’s a lot to qualify here. Learning isn’t a cut-and-dried matter of having a monthly reading list, or a quota for MOOC certifications. Those things matter, but they aren’t all that matters.
The kind of learning you should be after is a healthy mix of theoretical and practical (with a moderate inclination towards the latter).
Staying familiar with theory keeps you sharp, and grounds you in the logic underpinning your algorithms, models, and programs. Jobs get repetitive, and this is especially true for jobs you’re good at; beat the call of complacency by paying mind to the innovators at work.
Conversely, theory is at its best when it’s paired with positive action. There are many so-called data scientists who find their self-assurance in what they think they can do, and neglect to prove that they can actually accomplish anything. If you’re a data scientist looking to prove your value in the corporate world, your best shot is to keep yourself busy with hands-on projects.
You Do the Dirty Work
Data scientist is a sexy job title, but a good chunk of that work is far from glamorous. Before anyone can save millions by following your brilliant suggestion, odds are you’d have to make sure all your sexy data is in order.
Data is a lot more marriage than sex: before you can have all the good bits, you need to put in the work (read: the sometimes difficult, bang-your-head-on-the-wall kind of work). That includes deduplicating entries, turning numbers to string, separating first names from last names, doing fuzzy matches, and a bunch of other unsexy things.
It’s a long winding road from data to insight. The question is, are you willing to stick it out for better or for worse?
Remember, being a data scientist doesn’t exempt you from being a data janitor.
You aren’t Obsessed with the Title
If you’ve lived in the Philippines for an extended period of time, then you know that people get defensive about their titles. There’s no force in the universe that could absolve you for calling the wrong person mister instead of attorney.
As a rule of thumb, the kinds of people who obsess over titles are likely to tend to their ego before tending to their clients. It’s a distinct form of arrogance that stunts growth. After all, what mediocre fool would have the gall to suggest that a bona fide data scientist do better?
A good data scientist doesn’t need the title to do good work with data science. They need the tools for the job, and they need the drive to see a project thrive. They know full well that leaving a trail of happy clients in their wake says more than a title ever could.
It bears mentioning that there are whole discussions we’ve glossed over in the course of this article. There might be One Best Programming Language out there, and a seamless model is still a thing of beauty in the end.
However, the growing hype behind data science means that anyone who calls themselves an expert is inherently responsible for their clients’ outcomes. Whether we like it or not, data professionals are shaping the image of an entire industry in its relatively early days.
As such, a good, long look in the mirror can save you a lot of folded contracts and burnt bridges. There are habits and routines that every professional is at risk of developing over the course of a career, so don’t kick yourself if you fall short of Employee of the Month material.
Stay practical, learn as much as you can, and for God’s sake, take a page from Kendrick.