3 Valuable Perspectives that Data Scientists Need for Lifelong Learning

• Read original article here

As a recent data science graduate, I’ve been feeling a healthy dose of graduation anxiety. Swapping what is familiar for the uncertain. Handling the high-pressures of work. Figuring out “what will I do with my life?”These are some of the themes, but I want to talk about none of them though they’re all important. The salient thread in my graduation anxiety is far less tangible, much longer term, and astoundingly overlooked by ex-students: the way in which we think about and approach lifelong learning.

To make meaning of the junction that I’m crossing, I decided to describe three perspectives that have fundamentally changed the way I think about learning. None of the takeaways are technical in nature, let alone specific to any content area. This is by design. Data science professionals face an abundance of “what” to acquire through a career, and the selection of tools and techniques will constantly evolve. What is constant — and harder to come across — are the high level perspectives we must bring to bear through the ins and outs of our continuing data science education.

I used to humor myself with the notion that the only thing separating great data scientists from the mediocre was one’s adeptness at Googling their problem. The core of the joke (and what bears semblance with the real world) is this: Great questions trump great evidence. Roger Peng writes about this when he describes the pitfall of trying so hard to optimize an approach that the data scientist neglects to frame the proper question. It’s a blind spot that sounds an awful lot like golden hammer bias and happens in large part because data scientists (myself included) favor the shiny and complex tools, which unfortunately do more to obscure rather than crystallize the problem.

I’ve seen the same “evidence-over-questions” phenomenon in business school. We tend to reach for the same generic frameworks: a matrix for segmenting the competition, a tree diagram for growing profitability, a pricing model for showing elasticity of demand. All are great at simplifying the exterior world, but their application is much more nuanced in practice. What is required, I am convinced, is an exceptionalism in the skill of “asking”—relating, exploring, and articulating the hidden need of an individual or portion of the population. While I’m grateful that I walk away from graduate school with so many edifying frameworks, the greatest value of my education has been in learning what it means to harness the surprising power of questions.

There are two remarkable things about asking great questions: Having the creativity to formulate them and having the courage to ask them. So what is one thing we can do to ask better questions today? When I look back on my studies, Design Thinking was the single most valuable thing I did to improve the quality of questions I asked. While there are a lot of great nuggets in the Design Thinking field, the big idea is empathy.Empathy enables you (an outsider) to get in close proximity with the person for whom you are designing a solution, whether a customer, client, or colleague. Start by familiarizing yourself not just with the types of questions that are asked, but the way in which practitioners adhere to the design thinking process.

To step through the design thinking process is to encounter the unencumbered art of asking questions. I cannot count the number of times I have had to reject the temptation of believing, “I get this” or “I have the solution.” It is one thing to knowthat you need questions, but a completely different thing to bring discipline and patienceto the question process: the former only scratches the surface while the latter unearths deep insight. Discipline comes in the form of iterative interviewing when you’re continuously improving and sharpening the questions being asked. Patience comes in the form of steadily digging into the “why” behind each observation. The result is a profound understanding of the stakeholders and the context in which they think, feel, and behave. It is the sort of understanding that generates actionable insights in the present and unlocks higher forms of learning for the future. We get there not through rote application of algorithms and frameworks, but through a heightened sense of the kinds of questions we need to ask.

It’s like my former professor used to always say:

One resolution I made two years ago was to become great at telling stories from data (the strength I wanted to sharpen) rather than building great predictive models (the weakness I chose to set aside). When I look back on the evolution of these skills, it turns out that I was on to something with my data storytelling aspiration: a client is delighted by the insights I recovered from their data, a professor compliments my presentation of a commercial opportunity, a student mistakes me for a professional facilitator. It is easy to see the strengths narrative here: Innate Strength × Grit = Superpower. What is less visible is the pair of conditional variables that make strength your weakness and weakness your strength.

The first variable is ego, and it warns that our best attributes are also our Achilles’ heel. I once participated in the USAFacts Data Visualization Challenge, a school-wide competition that seemed right up my alley and that, if won, would grant the opportunity to meet one of the world’s most influential business leaders, former Microsoft CEO Steve Ballmer. With four classmates, I invested hours into finding the right story in the data — and if I’m being honest, there was a bit of hubris permeating:Here I am, a data scientist in training, competing against a few dozen Excel junkies in an arena full of data.It would have been a lie if you heard me say I never expected to place high.

The next day, during a practice session, my team’s visualization was nominated by professors and classmates as an exemplary exhibit of “What Not to Do” in the competition: it presented too much information, contained excessively flashy distractions, and failed to clearly communicate the intended message. Over-confidence had undermined me. I need no more persuading that it will get me again if it goes unchecked.

The second variable is authenticity, and it suggests that you couple your weaknesses with transparency. Now, I’m not saying go scream your weaknesses from the rooftops in the spirit of vulnerability. What I have learned, however, is to be open with yourself and with your colleagues about the areas you wish to improve. There are two things to remember about weaknesses:

It had never occurred to me that I could one day become “better than average” at what I am inherently awful at, but others before me have found paths to such step change. In the last week of school, an alum returned to class to share experiences from serving as CEO and Chairman of a Fortune 50 company. He opened with a 20-minute reflection of how mediocre of a student he was, even describing with elaborate detail the memory of an abysmal class presentation that still haunts him decades later. All this to emphasize that glaring weakness today (management communication, in his case) need not block you from becoming the captain of your industry tomorrow. Here was a person with impressive credentials, who spoke so eloquently in the moment, yet ceded such a different picture of his past. The dissonance was so intriguing to me that his advice formed a concrete lesson about lifelong learning in my mind. To paraphrase his words:

There is a common case in operations management that goes something like this:

Want cheap and fast? Go for McDonald’s, but don’t expect Michelin Star quality. The model reminds you that every endeavor requires you to weigh the trade-offs and make a decision. After thinking about all this, I was inspired to draw the following diagram for my own personal predicament:

The goal is to be good at all three. But how was I supposed to choose between highly sought text analytics training and highly practical small business consulting? Would I be okay skimping a few nights with my sons in exchange for a couple more hours on a Data Visualization competition? The answers may seem self-evident, but for me, they were a frequent occurrence and common source of confusion and regret.

Choosing learning priorities as a graduate student is no different from choosing work priorities as a professional. It’s really hard. The data scientist facing myriad requests must balance what is novel, what is important, and what is actionable to the company. The general manager in crisis must know what decisions to make now, what to make later, and what to delegate to others. If Successis enabled by Learningand Learningis governed by Time, then by what power is Timecommissioned? I knew nothing but that my own disappointment served as some indication that I hadn’t thought deeply enough about the things I most valued and the ways in which I might align them.

In my search for answers, graduate school provided an unexpected compass: Literature. In Aristotleand Benjamin Franklin, I discovered guidance for creating better learning habits. In contemporaries like David Brooks and Clayton Christensen, I found blueprints for defining and measuring my success. Shortly after, I started dedicating an hour every day to reflection, where I would enumerate daily thoughts and activities in order to discern what was important enough to keep and what was okay to let go. Attempting all this in graduate school at first felt like pool diving with only a faint idea of how to doggy paddle. Gradually, however, the life vest emerged in two discrete arrivals:

Clarity of purpose often provides the greatest return on time invested, especially in matters of work and life. While you certainly don’t need a masters degree to find your purpose, there’s something to be said about focusing your “why” and determining your principles during an extended period of hardship. In my case, graduate school did not define my purpose, but it was a wonderful forcing mechanism for unveiling it.

The second idea I discovered is called creative recombination. I used to think that managing myself was about dutifully using robust planning tools — creating master lists, working off the calendar, and syncing a virtual assistant. It wasn’t long until I realized that these tools are better suited to eking out small time improvements than for dreaming up ways to reshape self-management. The game-changing strategy for me was born of repurposing a technique from the study of innovation, which asks “What do I have lying around, and what opportunities can I capitalize on by merging them?”

In my final quarter of graduate school, for example, I started by clarifying the learning outcomes I needed to pursue: Be able to design and implement a range of machine learning algorithms in a real, applied healthcare problem. It sounds simple, but having that focus was a big reason I discovered what “recombinants” were available within my ecosystem: the professorwho linked me to the physician, who intimately understood the business problems faced by the hospital, which could provide large amounts of data. The business and data science elements were already waiting, and when merged together, they carved out a learning milestone that otherwise would have been much more difficult and time-consuming to pursue separately. I just needed a bit more focus and creativity. It was a powerful wake-up call, and one that I’m confident will go on to motivate and sustain my data science learning for years to come.

So What: Define your purpose, develop guiding principles, and stick to them.

Today, there is no shortage of resources to learn data science on your own. While this abundance can elevate efforts for lifelong learning, it also can erode them in three subtle ways:

So, before you dive in to that new data science lesson, remember that accelerating the learning curve is not just about collecting techniques in an additive way. It’s also about compounding the effect of learning with higher perspective. We move in that direction by shifting our attention to (1) more nuanced understanding of the problems we face, (2) sincere openness with teammates, and (3) a clear sense of why we’re choosing to dig into an idea.