Page 33 - The Real Work Of Data Science Turning Data Into Information, Better Decisions, And Stronger Organizations by Ron S. Kenett, Thomas C. Redman (z-lib.org)_Neat

18 The Real Work of Data Science

Manager: “I’m not certain. Maybe.”
Redman: “OK, your number is 25. Let’s talk again in two weeks and we’ll figure
out what to do next.”

To complete the story, the manager came back two weeks later, having completed 20 (of the
25). In 19, he had found a significant error. Some errors looked similar (the manager did not
yet understand “common cause”). He concluded that there was little doubt that the “overall
process was broken.” Subsequent work should focus on understanding the larger implications
for his division (a far larger organization than the group he managed) and sorting out what to
do about them.

Understanding the Real Problem

Note the contrast between the original need and context and those arrived at during the first
meeting. This example provides an object lesson in understanding the real problem. Data sci-
entists simply must engage with “customers” in their languages and talk through the apparent
problems to discover the real ones.
Frankly, with two warnings, we find that many people make this more difficult than it needs
to be. Three points to think through. First is our choice of the word “customer.” We find that
viewing decision‐makers as customers humanizes them. A customer can come from a differ-
ent part of the organization, be way up in the hierarchy, or be based in another part of the
world. But customers are people, with strengths, weaknesses, hopes, and fears. They are just
like us in that respect.
Second, “in their languages.” Just as we don’t want to get into the technical specifications
about how our car windshields were manufactured, so too we should not expect customers to
engage with us in the (to them) arcane languages of statistics and data science. We should
encourage them to speak their languages and make every effort to learn that language. This is
hard work – after all, those who drill for oil, run social media campaigns, and hedge securities
have developed their own specialized languages that are no more familiar to us than data sci-
ence is to them. So don’t be afraid to ask questions and to say things like “let me make sure
I understand.”
Third is “discovering the real problems.” It is a rare decision‐maker indeed who can articu-
late the real problem in the first try. Articulating problems is hard work, and there are so many
special cases, external factors, and political considerations that may cloud one’s head. It is the
data scientist’s job to help clear away the clutter, draw your customer out, and propose various
options. It is tempting to rush this work, but don’t. After all, as Albert Einstein observed, “if I
had only one hour to save the world, I would spend the first fifty‐five minutes defining the
problem, and only five minutes finding the solution.” 2
The first warning involves bad intentions. Some decision‐makers are not above using data
science to justify decisions they’ve already made (more on this in Chapter 11), to made others
look bad, or to promote personal agendas. So it also helps to develop a keen sense of smell – if
something smells bad, it probably is. You are unlikely to fully prevent such behavior, but you
must inform your boss and proceed with caution.

2 http://www.azquotes.com/quote/811850.