The main problem we encountered with this data set was the enormous number
of instances. In the final analysis, we ended up using less than one percent of all
the data that were available to us. It is very possible that the trends found in the
analysis are simply the result of the points that we selected from the entire data
set. Some potential solutions to this problem include either upgrading the system
to handle a larger number of polygons or creating a series of graphs using different
randomly selected data points.