Overview

About Analytics Framework Service

The Predix platform provides the Analytics Framework service to simplify developing advanced business analyses and deploying them to business operations. This service provides the following features for developing and embedding advanced analyses in industrial operations.

Analytics Catalog
A software repository for sharing reusable analytics, both subscription analytics purchased from Predix Catalog and non-subscription analytics you develop. Analytic developers upload their custom analytics and supporting files to the Analytics Catalog for easy management and reuse. The catalog also facilitates deploying analytics into production on the Predix platform.
Orchestration Runtime
An orchestration is a group of analytics to be run as a single unit. The orchestration runtime provides the ability to configure, execute, validate, and monitor analytic execution. You can perform analytic orchestration through the runtime's configuration and parametrization capabilities, reducing the need for custom coding and point-to-point integrations. As business needs evolve and new analytics are developed, these configurations can be readily updated and redeployed. The runtime is an efficient, scalable, cloud-based approach for both the development and production use of advanced business analyses.
Scheduler
The scheduler provides the ability to schedule the execution of analytics or orchestrations of analytics on time-based intervals, also called a job. You can create a job, retrieve both job definitions and history, update definitions, and delete jobs.

Based on standard architectures, open technologies and methodologies, the Analytics Framework provides the following value to your organization.

Analytic Developers
  • Faster analytic solution development by abstracting the complexity of developing and deploying orchestrations in the cloud.
  • Ability to test and validate orchestrations in the cloud.
  • Faster analytic solution deployment by favoring configuration over coding. Changes to orchestration configurations (data sources, asset applicability, and so on) have been simplified.
Application Developers
  • Faster analytic solution development by abstracting the complexity of developing and deploying orchestrations in the cloud.
  • Ability to test and validate orchestrations in the cloud.
  • Faster analytic solution deployment by favoring configuration over coding. Changes to orchestration configurations (data sources, asset applicability, and so on) have been simplified.

The catalog supports analytics written in Java, Matlab, and Python and the runtime supports analytics from the catalog.

The following figure shows an example workflow as an analytic moves through the framework, from development to execution. The analytic moves through the process starting with being uploaded to the catalog, then tested, and then deployed to a production environment. Next, the orchestration runtime runs an individual analytic or an orchestration of analytics for evaluation or production use.

Figure: Example Developer Workflow Using Analytics Framework Service


Stage 1: Develop the Analytic
Develop the analytic in either Java, Matlab, or Python. This work is done outside of the Analytics Framework service.
Stage 2: Upload the Analytic to the Catalog
Add the analytic to the catalog and upload the artifact.
Stage 3: Validate the Analytic
(3a) Test the analytic with a sample data set and verify the results.

(3b) If necessary, modify the analytic and upload new artifact to the catalog.

Stage 4: Deploy the Analytic
Deploy the analytic to your production environment.
Stage 5: Configure the Orchestration
Configure the order of analytic execution within an orchestration workflow and the mapping of analytic inputs and outputs to data sources.
Stage 6: Execute the Orchestration
Stage 7: Monitor the Orchestration Execution Status
Retrieve the orchestration execution results for each orchestration step.
Stage 8: Schedule a Job to Run an Analytic Execution
Schedule the execution of an orchestration or an individual analytic on a time-based interval, also known as a job.