Analytics

Purpose

This article will provide instructions on how to create, modify, search and delete Data Models on the Process Design Studio tool.

Definition

The objective of Analytics is to predict a certain aspect of an Event’s data, based on a machine learning tool that will use Data Warehouses as source for analysis.

The result of a Data Model will be used on the comparison that will initiate a Workflow.  This is set up on the Rule Builder tab of the Workflow Event´s configuration.

Create

In order to create a Data Model, click on + New Model.

Settings

Add a name to your Data Model.

Select the Data Warehouse Table that will be used as the source for the data analysis.

• All fields from that Data Warehouse will be included on the list and cannot be removed.

Select the Data Warehouse field you would like the service to predict.

You can add a description to your Data Model if needed.

 Select the Problem Type you would like this Data Model to use. Problem Types are the different analysis methods that can be used for this machine learning functionality.

• Regression: Estimates the field to predict based on one or more other variables or attributes that are correlated with it.
• Binary classification: Assigns an individual to one of two predefined and mutually exclusive classes based on their attributes.
• Multiclass classification: Assigns an individual to one of several classes based on their attributes.
• Automatic: This allows the service to automatically identify the Problem Type needed for that specific Data Warehouse.

 You can associate the Data Model with one or more Business Processes.

 In order to get access to the Jobs and Endpoints tabs, you must create the Data Model first.  So after all previously mentioned fields have been completed, click on Create.

You will be redirected to the Predictive Analytics page.  In order to keep on configuring the Data Model, find it on the list and click on the pencil icon.

 When accessing a created Data Model, all three tabs will be available.

Jobs

A Job defines the analysis criteria over the Data Warehouse previously selected. Each Data Model can have different Jobs.

 The Job will already have a name and number of candidates pre-established, and the Type of Problem will be taken from the Settings tab, but all these details can be modified by clicking on each field.  Click on the name box for editing the name.

Click on the Type of Problem field to change the learning model.

Click on the Number of Candidates field to modify the number of calculations to be used.  The maximum number allowed is 250.

The Status column will indicate if the Job is:

• Pending: When the Job has not been run yet. 

• Running: Once you click on Run Job and the process is starting. 

• In Progress: When the Job is being processed.  

• Completed: When the Job has been processed and the Endpoint is ready. 

At the end of the line there are two icons on the same box:

• Trashcan: Click on this icon if you would like to delete the item.
• Run Job: Click on this icon whenever you are ready to execute the analysis.  You won’t have an Endpoint unless you run the Job.  If you click on Save without having ran the Job, it will disappear from the list and will have to be created again.

Endpoints

An Endpoint is the result of a Job.  On this tab you will see the Name, Creation Time, Type of Problem used, Status and Accuracy of each job ran inside the Data Model.

Once a Job is completed, you should click on Deploy in order to make its URL available for the comparisons that will be set on the Workflow Event.

 Once the Job is deployed, the status will change and the option for deletion will be available.

Modify

In order to edit a Data Model, click on the pencil icon to the right of the item on the list.

Search

Coming Soon

Delete

In order to delete a Data Model, select the item on the list and then click on the Delete Model button on the top right of the screen.