Database Reference
In-Depth Information
A Step in the Process
How is “communicating data” distinct from the other steps in the overall process that begins
with a question and ends with a shared insight?
Figure 1-1
presents the overall data discov-
ery process, and shows where communicating data fits in that process.
Figure 1-1. The data discovery process
The highly iterative process often begins with a
question
, which can be specific (“which
combination of products occurs the most often?”) or general (“what can we learn about his-
torical sales of our products?”). The next step is
gathering data
if it's available (e.g., historic-
al sales). Then comes the often arduous process of
structuring data
, also called “data
munging” or “data wrangling.” In this step, data is formatted, shaped, merged, converted,
and otherwise manipulated into a form that is amenable to the next step,
exploring data
. In
this step, the data is viewed and analyzed from a number of angles until one or more insights
are gleaned. These insights form the message involved in
communicating data
, the step at
which quantitative statements are shared with others. While this topic primarily concerns this
final step, it will also touch on the other steps in the process, as they contribute to the forma-
tion of the message to be communicated.
In order to examine the idea of communicating data in greater detail, let's return to the birth-
place of information theory: Bell Laboratories.
A Model of Communication
The year was 1949, and two employees at Bell Laboratories—Claude Elwood Shannon and
his coauthor Warren Weaver—published a seminal article in the University of Illinois Press
called
The Mathematical Theory of Communication
. In it, they introduced a model of com-
