Page 47 - 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

34 The Real Work of Data Science

With this in mind, take two steps. First, provide your explanation of how to read the chart
in the notes page of your PowerPoint or slide deck program. Second, annotate the graphic, as
done in Figure 7.2. Although annotations do not replace a well‐told story, they do give the
reader some inkling of what is involved.

The Best of the Best
Finally, to most decision‐makers, an ounce of insight is worth a ton of analysis. Thus, one out-
standing graphic that cuts to the heart of the issue at hand and guides next steps is worth more
than hundreds of mediocre ones. Seek that graphic. Presented this way, data science is raw
power!
There are dozens of books on this topic, and all good analysis packages help you develop
amazing graphics. We’ve already cited Tufte. Other examples follow.
Hans Rosling (2007) developed bubble plots to show how phenomena evolve over time. He
claims that preconceived ideas inhibited academics from properly realizing trends in demo-
graphics and child mortality statistics. His presentations showed that, in this respect anyway,
third‐world countries are not third world at all.
Jean‐Luc Doumont (2013) trains scientists to communicate more effectively. Among other
things, he highlights the pitfalls in communication with PowerPoint.
Jorge Camoes (2017) provides 12 suggestions for effective thinking with data visuals, with
examples. His main points are the following:
• Data visualization is not enough; you have to have the contextual knowledge to detect and
interpret patterns.
• Visualization fails not because there are too many data points but because the presenter
doesn’t understand the data or does not care about the message.
• Simplicity is not minimalism or removing junk. Not surprisingly, he advocates removing the
irrelevant, minimizing the accessory, adjusting the necessary, and adding the useful.

Implications
The practical reality is that few decision‐makers understand, or even care about, p‐values, sig-
nificance tests, and the like. Nor do they appreciate how hard you work to tease insights from
the data. But they do care about results, particularly those presented in simple, visual, compel-
ling ways that are directly applicable to the problem at hand. Thus, the real work of data sci-
entists involves telling compelling stories based on your insights, providing simple graphics
that drive the message home, and helping decision‐makers understand the full implications
and context. Data scientists need to get very good at this.