Visualization Strategies for Exploring Large Datasets - Data Just Right: Introduction to Large-Scale Data and Analytics - page 88

Figure 7.2 Charles Joseph Minard's 1869 work Carte figurative des
pertes successives en hommes de l'Armée Française dans la campagne
de Russie 1812-1813, described by Edward Tufte as the best statistical
graphic ever drawn
Both of these stories illustrate not only the power that visualizations can provide
but also the complexity and care that it takes to condense numerical data into a com-
pelling work. Each of these data visualizations took a great deal of multidimensional
information and made it accessible. In a simple two-dimensional space, ideas about
time, space, and metrics were condensed into a narrative. The impact of a great visual-
ization is that it not only tells a great story but also can be used to convince others of a
particular point of view.
Both of these masterpieces of historical visualization are inspiring, and the abil-
ity to create visual representations of data is more accessible than ever. Many people
are familiar with the graphing features of common business productivity tools such
as Microsoft Excel. Excel has excellent capabilities for creating charts and graphs, but
the push-button aspect of tools such as Excel means that it is very easy to generate
charts and graphs that are not particularly compelling or which tell a misleading story.
Because spatial representations of data can be so powerful, a good data scientist must
take great care to understand what types of depictions best support his or her narrative.
An overarching goal of information visualization is to communicate abstract con-
cepts using spatial features. Aesthetics are important, in much the same way that the
way one speaks or writes when telling a story can affect the recipient's experience.
However, one pitfall of data visualization is not being familiar with which types of
data can be presented using which type of visual representation.