This data set was the one of the largest data sets we encountered. Graphing all the
data points was impossible because of the performance limitations of the RAVE,
so only 3500 of the data points were plotted. To analyze this data set, we ran the
ChiSquared ranking algorithm to identify the five most important attributes. We
then created two separate graphs using variables of the same nature. For example,
one graph consisted of the data for shade at 9AM, 12PM, and 3PM representing
the x, y, and z axes respectively, with the size representing the slope of the terrain
and the color representing the elevation. To further analyze the data, each graph
was then split up by the cover type so there were eight graphs in all for each
analysis method. For this data set, we decided not to use the conventional display
environments which consisted of graphs being placed in a building of some sort.
Instead, we created a forest using trees made of cones and cylinders and embedded
the graphs in the midst of the forest. The graphs were set up in the new semicircular
arrangement to allow the user to easily view all the graphs by simply
standing in the center and turning to the left or the right. Instead of using gtb-axisl
to display the label, we opted to use a sign to display the variables represented by
the axes and the color and size of each data point