Figure 1

Aggregation scheme. Multivariate data is represented as a table of N rows and M columns, where each column can be of a different scale type. Rows and columns can be aggregated independently, resulting in hierarchies as indicated by the attached dendrograms. Aggregation of rows is applied column-wise, according to the respective scale type, while the aggregation of columns is only supported for compatible scale types. For example, the set of leaf nodes of v is a collection of rows R _v and the set of leaf nodes of w is a collection of columns. Aggregating v results in a new row with values determined using rows R₁, R₂, and R₃ (red square). Similarly, aggregating w results in a new column C _w with values determined from columns C₂ and C₃ (green square). As a result of collapsing both v and w, the overlap of the red and green squares will be represented by only one cell.