One of the key features of the dynamic parallel coordinate plot is the ability to adjust which variable is displayed on each axis. By selecting different variables on different axes, relationships between pairs of variables can be interactively explored.
In the climate example provided, the seven vertical axes are arbitrarily
set to the default variables; that is, from left to right, in order that
the variables were stored in the database (see below
for definitions of the variable abbreviations of the plot). However,
it is quite conceivable that two variables not adjacent to each other in
the database (say, sea level pressure change and humidity at 700 mb, may
prove to have a surprising or unexpected association which, using the PCP,
would be discerned only when the variables are moved to adjacent axes.
The figures at left detail this manipulation.
Variable abbreviations and explanation of data
The data displayed comes from an unsupervised classification of climate variables from model output of winter weather in the western Gulf of Mexico, Mexico, and Texas. More detail on the data and its sources can be found in the IJGIS paper.
The unsupervised classification returned 27 classes (labeled class 0 through class 26), the members of which are grouped according to some similarity in or among climate variables. Also returned is the probability that each observation belongs in the class assigned; for some observations, the choice of class is more certain than others because these observations are closer to the most typical member of the class to which they belong.
The space-time-attribute variables comprise the remaining five variable choices in the example dynamic PCP. The data consist of three days' worth of model output for 20 grid points, arranged in a 4 by 5 rectangle (left). The days are plotted in the sequential_date variable; the numbers (31352-31354) correspond to the number of days since an arbitrary start time, in this case, Jan. 1, 1900. The attribute variables are sea level pressure change (sea_level) over the 24 hours prior to the observation, precipitation in cm, and humidity (water vapor pressure) at 700 mb (mid-level in the troposphere).
