This demonstration shows a few visualizations for Data.gov's Dataset 1612. More information may be found on Tetherless World Constellation's data-gov wiki. Please note that some of the visualizations require Flash. By Tim Lebo.
The following parallel coordinate system shows the number of United States military service members in 2009 with and without children, broken down by their pay grade and whether the service member is single or married. Each blue line depicts a group of military members who share the same pay grade. Unfortunately, the lines cannot be selected to provide identification of the demographic depicted, nor can they be highlighted to facilitate reading a single line through all coordinates . The potential benefit of this feature is clear when observing the crowded group at the bottom of this parallel coordinate system. Another issue is that http://code.google.com/p/parallel-coordinates/ cannot render more than one parallel coordinate system on a single page, so this is an image. Being able to color the lines differently would allow us to distinguish between the classes of pay grade (enlisted, officer, warrant officer), but, unfortunately, only one color is permitted. Control for ordering within the Pay Grade coordinate is also not provided, so O-10 is (lexicographically) between O-1 and O-2.
When choosing the smallest demographic as the invariant, the motion chart is less than optimal. This is because only one quantitative variational component is available in the data, while the motion chart can convey three in a single instant of perception (using the visual variational components x position, y position, and color). This sub-optimality is seen by the y=x line that the colored circles form. Because the data represents one time sample, the animation through time is also not available. The abundance of nominal and ordinal variational components (Children, Gender, Marriage Status, Pay Grade, Service) can be seen in the "Color" pull down. The numbering shown when brushing over a colored circle is a meaningless name for the smallest demographic; it represents one of the combinations of every demographic characteristic.
By changing the invariant to service (Army, Air Force, Marine Corps, Navy), we earn 8 non-temporal quantitative variables:
Unfortunately, so may quantitative variables makes it difficult to choose which pairs to choose for the axes. Parallel coordinate systems can show more than two in a single instant of perception. It is clear that the Marine Corp has fewest of all groups, since the line continues across the bottom. The Army does not share the same correlation that the Navy, Marine Corps, and Air Force share with the first four dimensions. The Army and Air Force swap dominance for service members married to other service members.