Original article was published by Elena Marocco on Artificial Intelligence on Medium
A question has been running through my head over the past few weeks: “How can I teach data science better?” What I didn’t realize until later, however, was that I was also asking,
“How can I be a better leader?”
This issue was on my mind because last week was a learning week at the company where I work, Evo. That means that every moment not spent addressing urgent client needs was devoted to learning initiatives. This time around, I was continuing to develop my first data science base-course for Evo University, a soon-to-be-released set of public courses on data science concepts related to the work we do.
While I’d long been teaching some of the concepts touched on in my digital course to new members of my team or clients interested in understanding the model better, this was the first time I’d set out to create a formal course on the subject. It’s been an exciting challenge, one that has reinforced key lessons that apply to more than simply instructing others on data science. I’ve learned to be a better data science leader. Here’s how.
1. Assume nothing.
When you are building an introductory course, your students don’t have a broad knowledge base to build upon. You have to give them the basics in the course itself or you will lose them. When I started structuring my course, I struggled to figure out what skills and vocabulary I could infer my students already had developed.
That’s when our training expert advised me to simply “assume nothing” and define any technical concepts or skills referenced within the course. After that, creating my course went much more smoothly. I could feel confident that my students understood the basics needed to move on to the more important topics. Eliminating my assumptions introduced a desperately needed clarity into the process.
This refrain isn’t only valid when teaching; it’s just as critical when leading. Any time I’ve assumed that my team will understand what I need them to do without clearly laying it out in a ticket or explaining myself directly, we get poorer outcomes. While I’d long dedicated myself to precision, this simple phrase drove me further.
When you assume, you fail to provide people with what they really need.
A good leader asks first and then provides answers that empower others to accomplish their goals. Never assuming is a good policy when you are developing curriculum — and critical to leadership.
2. Build little-by-little from the basics.
The more I teach data science and other technical concepts, especially to people without a technical background, the more I recognize how important it is to start small. Unless you introduce the topic at a scale anyone can easily engage with, you’ll lose their attention before you ever get to what you want them to understand. Theory underpins effective learning. You can’t understand more complex topics until you have mastered the basics underneath them.
Once the theory becomes clear, you can add layers of complexity to them to build to more difficult concepts. Little-by-little you help your trainee achieve mastery of the topic, building slowly to the most critical information. Whether I am trying to impart knowledge or simply trying to give my team a new task, I have to start from the most basic elements and build slowly to the more complex ones. An empathetic teacher is careful not to overwhelm their students, just as a good leader is careful to never ask too much of their team at one time.
3. Make all information actionable, supported by visuals and examples.
When it comes to data science, theory is important, but knowing all the theory in the world is useless unless you can apply it. Application is the critical step.
After all, if you can’t apply what you’ve learned to real-life situation, have you actually learned it at all?
I studied math, so I’d never suggest that theory isn’t important; it’s taken me far in my career. Nonetheless, information needs to actionable for it to be most useful to your students — and your team. Whether I’m teaching data science or communicating with my team, the information has to be relevant and tangible so they can act upon it.
How? Using as many examples, images, charts, videos, and other visuals as possible. They say a picture is worth 1,000 words. The truth is that any medium that helps contextualize what you’re learning can become a powerful tool to enhance understanding.
As a data scientist working to make AI more accessible, I have seen first-hand how data visualization skills can transform a complicated topic into one that anyone can understand. Visuals, cases, stories, and interactive examples make your argument more powerful because they make it more concrete. It’s difficult to make the leap from theory to your own example, but it’s much easier to relate two examples to each other. Whether learning or following, actionable information is far easier to envision and use successfully than either extreme from vague generalities or highly complex details.
4. Remember what it was like when you were starting out.
Perhaps the primary lesson emphasized last week — and one we all have to learn again and again — is to put yourself in your audience’s shoes. Just because you know exactly how to do something, it doesn’t mean that knowledge will come as easily to someone doing it for the first time. You must have an empathy for them as beginners and support them how you needed when you were still starting out yourself.
That empathy is easy to conjure when you remember what it was like before you were the expert.
Every teacher started out as a student; every leader started out at the bottom.
Reach back to what it was like before to provide the kind of information and support you had (or wish you had). An effective teacher addresses each student’s unique emerging needs. An effective leader does the same for their team.
Effective leaders are empathetic teachers
When I set out to become a more effective instructor, my only goal was to create better training for people who wanted to master key data science concepts related to my work. I was excited about making these courses both instructive and actionable.
What I didn’t expect, however, was that reinforcing these lessons could improve my data science leadership. A leader is often a teacher — and even when that’s not my role, these four lessons still improve leadership skills. No matter what kind of role you plan to take in your career in data science (or even far from our field), the skills of an empathetic teacher can help make you a more effective leader.