If you are using IBM SPSS Modeler and you are regularly working with data sources to build models and other analytical streams then this node will improve your productivity and it will enable you to work more efficiently with test, production and deployment streams. Enhanced Metadata for IBM SPSS Modeler provides a new Metadata node which addresses a number of the limitations of the existing Type node:
- Valid values for ranges, sets and flags can be defined on one place and then mapped to multiple fields. By updating one metadata definition, the metadata for all fields using that metadata is instantly updated
- The Metadata node editor dialog allows multi-line text to be pasted directly into set values and value labels making it simpler to transfer values from other sources
- The documented Metadata node scripting interface allows valid values to be defined and mapped to multiple fields more simply than would be the case with the standard Type node. For example, a script could read value definitions from an external file or REST call and update the metadata definitions each time a Modeler stream is run
- A single Metadata node can define the metadata for the whole Modeler stream. Other Metadata nodes can consume that metadata, ensuring there is a “single source of truth” for the whole stream and that consistency is maintained automatically
- Many standard data representations are available as built-in types within the Metadata node. For example:
- Common range types such as probabilities between 0.0 and 1.0, percentage as 0.0 to 100.0
- Common integer encoded discrete values such as day of the week (as both 0-6 and 1-7) and months of the year
Other advantages of the ‘Enhanced Metadata for IBM SPSS Modeler’ include:
Enhanced Metadata for IBM SPSS Modeler is scalable to dozens or hundreds of metadata definitions.
The Metadata node is fully configurable using the Python scripting language provided by IBM SPSS Modeler. Online documentation for the node includes its scripting reference.
Who should use the Enhanced Metadata Node?
If you are using IBM SPSS Modeler and you are regularly working with data sources to build models and other analytical streams then this node will improve your productivity and it will enable you to work more efficiently with test, production and deployment streams.
- IBM SPSS Modeler 17.x or 18.x
- Windows 7, 8 or 10 64-bit operating system, or Mac OSX
- IBM and SPSS are registered trademarks of International Business Machines Corp.
- Microsoft and Excel are registered trademarks of Microsoft Corp.