Database Reference
In-Depth Information
7
Exploring Large Datasets
There is no such thing as information overload.
There is only bad design.
—Edward Tufte
A
ccording to Bit.ly's Hillary Mason, data scientists generally do “three fundamen-
tally different things: math, code... and communicate.”
1
Although some of the tech-
nologies available in the data toolbox primarily focus on engineering whereas other
technologies focus on mathematical analysis, the visual representation of information
requires a combination of both of these skills along with an extra helping of com-
munication skill. One could say that a goal of data visualization is to communicate
abstract concepts that emerge from the world of math and metrics using the more
human language of spatial representation.
The current practice of data visualization has a rich cultural history backed by
many decades of pioneering and practice that dates to long before the digital age.
Aesthetic considerations are important to mastering the communication of visual
information, so this is not a field that one can be expected to master in a short amount
of time. Although the field offers a great deal of time-tested best practices, innovations
are still being developed thanks to constantly improving interactive digital technology
and the practice of analysts sharing more and more datasets on the Web. The use of
techniques borrowed from fields such as user-experience design is helping visualization
researchers understand how to best communicate data narratives.
Although the world of data visualization is rich enough to fill a library of topics,
this chapter will touch on some practical considerations for dealing with large datasets.
We'll also take a look at popular open-source software that is often used for building
compelling visualizations for both desktop and Web.
Search WWH ::
Custom Search