Visualizing Cassie Kozyrkov’s Guide to Data Science Teams
Pablo Picasso said, “the problem with computers is all they can do is provide answers.” I think his message was: questions matter as much as answers.
For example, if you’re a business leader building a data science team, where do you begin? Cassie Kozyrkov, Chief Decision Scientist for Google, shared her top 10 roles in AI and data science. It’s an excellent place to start. Her list opens as many questions as it answers… Who do you hire first? Should you recruit social scientists before data scientists? How do you grow your team over time?
The art of analytics about questions, not only answers.
The newest Spotfire Spider Chart Mod by Arnaud Varin helps visualize a group of variables associated with a decision. So I decided to use it with Cassie’s list of data science roles. The skills are arranged in order of importance, clockwise, beginning at 12:00 (click on each image to zoom in).
I gave “Data engineer” and “Decision-maker” 10/10 in importance, 9/10 points for analyst, 8/10 for expert analysts, and so on. Her rankings look more like a nautilus than a web!
But when do I hire my team, and in what groupings? After a few readings, I translated Cassie’s list and suggestions in her text to form a step-by-step approach. The first two hires on her list are a data person and a decision-maker.
Let’s look at how our data science team should grow over four years, from that small initial team to form a fully-stacked organization. Here’s your hiring plan, year by year:
When I created this chart, I noticed “social scientist” is ranked #8 on Cassie’s list and falls into year four. That got me thinking… My mind went back to the early 2000s Wall Street when algorithmic trading exploded. The social aspect of data science was a considerable obstacle in the early days. Traders were slow to embrace algorithms in fear of their job. The ones that did engage transformed trading as we knew it.
This visualization sparked a new question. Should you hire a social scientist before a data scientist? I think so. The visualization didn’t just answer a question; it caused me to think about it.
The data fun is endless! Cassie warns to carefully vet Expert Analyst, Statistician, and Machine Learning skills during the interview process. These six spiders show how each interviewer rated a candidate on these skills. Catalina and Monica saw something nobody else saw.
The most compelling aspect of these visualizations is that they sparked new questions in seconds. Visual analytics becomes the sixth sense, a second brain that leads to new insights about the order of hiring, importance, timing, depth of skills… Picasso was right: it’s the questions that matter, not only the answers.
