However, we may be looking at a process that we do not understand enough to attempt to model and, for the time being, we may mainly want to ``look at the data''. Even this task is not very simple and requires some thinking. Tufte has written some beautiful books about visual display of quantitative information. The challenge is to produce visual appealing displays that uses our brain processing ability to convey the maximal amount of information. The fact that we ``process'' images has to be taken into account also to make sure that the display doesn't induce the observer to believe in patterns and effects that are not present in the data.
Microarray data analysts have taken ample advantage of color to display the numerical values of a matrix in an appealing way. Reordering of genes and varieties has also been used in these plots. Particular attention should be paid to the method used for the reordering. Patches of uniform color are visual appealing, but they are effective in representing the data only when the associations between genes and varieties that they suggest are real.
There are various methods to achieve a reordering of the data, some of which will be mentioned in the section devoted to clustering, other are referred to in the references. There are entire journals devoted to graphical methods of statistical analysis, so we really aren't aiming to be exhaustive. A particularly useful tool for graphical analysis of data sets is Xgobi.