Page 51 - 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

38 The Real Work of Data Science

Last, there is intuition. In the examples just above, both the marketing manager and the soft-
ware development manager could explain the soft data and the rationales they used to reach
their decisions. But even in the simplest, most straightforward situation, uncertainty remains.
Thus, all important decisions are made in the face of uncertainty – if for no other reason than
the future is unpredictable. And here intuition must take over!
Data scientists must embrace this reality, taking three steps: First, intuition should not replace
sound inferences based on trusted data. Rather, it should take over where the data leaves off.
They must help decision‐makers understand the distinction. Second, wherever possible, they
must quantify the uncertainty. And finally, they must cultivate their own intuitions.

Implications

Data scientists should view the various modes of inference as tools, just as they view R and
Hadoop. And they should learn to use, and combine, them all. Too many data scientists prefer
to stare at the data and do not think deeply enough about generalization of findings.
Thus, the real work of data scientists involves reasoning from the data to the situations of
interest to the decision‐maker. There are many ways to do so, and data scientists should
embrace them all. Data scientists should strive to remove as much uncertainty as they can in
their analyses and quantify (or at least clarify) that which remains. They must recognize that
no analysis, no matter how thorough, removes all uncertainty, as there are just too many things
that can go wrong or change. And they must develop their own intuition.