Principal components analysis is a method for transforming a given set of observed variables into another set of variables by determining the principal axis required to minimize the perpendicular distance between the data points and the axis (i.e. from that point perpendicular to the principal axis).