Data Format

The Multiple Factor Analysis platform uses a data table that contains column groups, or sub-tables. Each column group, referred to as a block, can have a different number of columns. Each block of columns can represent different measurements. The columns or blocks do not have to be in JMP column groups. However, the platform is easier to launch when the columns are grouped into blocks in the data table.

The data table rows represent the items that are being measured. Observations for each item must be in a single row. For example, Figure 7.5 shows a table that is measuring attributes of 16 wines from different vineyards. The column panel shows the column groups, or sub-tables, for each panelist. The Vineyards in rows 17 and 18 are not assigned a Region. The analysis could be used to explore which region the vineyards are most aligned to.

Figure 7.5 Partial View of a Data Table for Multiple Factor Analysis 

Note: Missing observations are replaced by the column mean. When missing observations result in no variation for a column, the column is excluded from the analysis. Missing rules are applied to all variables, including supplementary variables.