Principal Component Analysis (ViSta-PrnCmp) is a statistical analysis and visualization method for picturing the variance in a set of continuous multivariate data. Only the numeric variables in the data are used. 

The ViSta-PrnCmp visualization shows the maximum variance picture of the data: The first principal component is the linear combination of the variables that has the maximum variation. The second principal component is the linear combination of the variables that is orthogonal (at right angles) to the first one, and has the maximum remaining variation. The third is orthogonal to the first two, and has maximum variance. The scree plot is also shown.

There are two major decisions to be made:

Before the analysis, you must decide whether to obtain the principal components from the correlations or covariances of the variables. The choice of covariances can only make sense if your variables are all on the same scales. If they are not, you must choose correlations. If they are all on the same scales, you may choose either, but correlations is usually most reasonable. Choosing covariances means that the original differences in variance between variables will effect the results. Choosing correlations means this difference in variance will not effect the results.

After the analysis, you must to decide how many principal components are needed to adequately summarize the data. The interpretation for the model will help you make this decision.
