Scoring Outside the Box: Code Examples

In JMP 13.0, JMP added the ability to generate scoring code from its predictive models in three new programming languages: C, JavaScript and Python.

By tapping into the unique strengths and vast library ecosystems available to these languages, we hope to expand the possibilities of what users can do with JMP models and lower the costs of deployment to production.

This archive contains different applications built around JMP-generated scoring code. You will find examples of how to:

• Use JMP to publish a model to the new Formula Depot and generate scoring code in multiple languages.
• Compile your C scoring code into a dynamic link library (DLL) for execution by a command-line application.
• Run your JavaScript scoring code inside an interactive rich-client web application.
• Deploy your Python scoring code to a web server to provide scoring-as-a-service to a web client.
• Use a Jupyther notebook to load an Excel spreasheet into Pandas, score it with your converted Python JMP model, and visualize the results with Bokeh.
• Too much data to fit in memory? Execute your Python model in a distributed Spark cluster (experimental).