Wednesday, August 29, 2012By Andrew Disselkoen
Part of creating a good visualization involves knowing a bad visualization when you see one.
While style is subjective, some visualizations cannot avoid leaving the viewer befuddled. Here’s one example:
GE (data sourced from BP’s Statistical Review of World Energy, June 2010)
The only way to understand “Gas Giants” is by reading the text - the graphics lend nothing to understanding. Differing size boxes distort the comparison between countries. The chart intends to measure percentage of gas reserves within a country - one data dimension. Yet, the graphic changes across two dimensions (height and width). Naturally, the mind attempts to compare volume - but this is a more difficult comparison to make. Any comparison that could be made is confused by the arrangement of the boxes. The white space in the lower right-hand corner leaves the viewer wondering if the designer made a mistake.
In the end, the viewer must read the numbers to have any idea what the visualization attempts to communicate. Since the viewer cannot understand the graphic without reading the numbers - a simple bar chart or list of numbers would have presented the information much more clearly.
Stand Clear: The Most Used Subways
Tuesday, August 28, 2012By Andrew Disselkoen
Once and awhile, one stumbles across an interesting and effective visualization. Visualizations can often be too simple or too complex. Finding a happy medium can be difficult, but I think this graphic achieves that goal. This subway graphic is data-rich and displays multiple data dimensions. While the chart visually displays track length and ridership, a small geographic map details the track arrangement. A small table on the right displays information best shown in table form. This visualization works because it successfully plots two variables on the same axis. The viewer is easily able to compare and relate ridership and track length to infer the level of congestion when riding the subway.
Stand Clear: The Most Used Subway Systems in the United States and Around the World
Good Magazine Issue 15
Data Visualization Guidelines
Monday, August 27, 2012By Andrew Disselkoen
Charles Joseph Minard 1969
Modern data visualization often focuses on powerful computer-generated graphics and highly saturated colors. However, one of the most renowned statistical graphics was created in 1861 by Charles Joseph Minard. Depicting Napoleon’s Russian campaign of 1812, Minard captures six variables in a coherent graphic. The width of the line depicts army size with the black line denoting his retreat. The lines are arranged on a two-dimensional map showing military movements and direction. Lastly, temperature is plotted along the bottom exposing losses in the harsh weather. Free of computer-generated graphics and bright colors, Minard demonstrates form and substance are two of the most important elements of a good data visualization.
Tips for Effective Graphics
Show, Don’t Tell
Induce the viewer to think about substance rather than methodology. Printed type should be used sparingly and free of abbreviations and cryptic codes. The visual should be the primary means of communication. If information cannot be communicated clearly and creatively using visuals, the graphic becomes an unnecessary addition.
Avoid unnecessary graphical embellishments and bright, gaudy color schemes. Color palettes should be cohesive and calming. Graphics should primarily communicate the message without unnecessary embellishment. Colors, patterns, and graphics should complement the design rather than compensate for poor visualization design. Computers have enabled a spectrum of customization, but embellishment cannot overcome poor design and planning.
The purpose of a data visualization is to display data in a clear and unique way. Almost everyone can create an excel chart. Good data visualizations incorporate multiple data dimensions. The more information that can be cleanly displayed in a small space, the better. Don’t insult your viewer with a simple graphic, but also don’t confuse with an elaborately cryptic design.