Cluster Analysis for Multiple Data Types
Can the Clustering methods available in SPSS handle mixed data types, i.e. a combination of categorical and continuous cluster variables?
Resolving the problem
The TwoStep Cluster procedure, which was introduced in SPSS 11.5, can use a mix of categorical and continuous predictors. The categorical cluster variables may be numeric, short string variables (8 or fewer alphanumeric characters), long string variables (more than 8 alphanumeric characters), or any combination of these variable types.
TwoStep is included in the SPSS Base module. It can be accessed in the menu system from
In the main "TwoStep Cluster Analysis" dialog, you can paste cluster variables from the variable list on the left side of the dialog to either the "Categorical Variables" or the "Continuous Variables" on the right side of the dialog.
You can find a tutorial for TwoStep Cluster under Help->Case Studies. Expand the link for Base System in the Table of Contents that appears and scroll down to the TwoStep Cluster link