Creating metrics for Data analysts

IBM watsonx BI uses metrics and the underlying enriched metadata that is used to create metrics to answer questions and provide insights. You can select and prepare data for conversations in the Create metrics workflow.

The Create metrics workflow is where you connect to or upload data, data undergoes metadata enrichment, you can then generate metrics and visualizations automatically or build them manually, and model data to create joins, relationships, and calculations.

While both Analytics consumers and Data analysts can create metrics, only Data analysts can specify access permissions, and publish metrics and their visualizations to the Metrics catalog. Data analysts can assign published metrics to users in an organization from the Metrics catalog. These assigned metrics display in the users' Key metrics.

After metrics and any associated visualizations are published to the Metrics catalog, Analytics consumers can browse the catalog to see the metrics they can ask about and pin metrics to their Key metrics panel. They can also ask questions against a specific metric directly from here.

As a Data analyst, here is an example of the steps you can follow to create metrics, publish them, and assign them to users in your organization.

Prerequisites

Before you create metrics, familiarize yourself with the key concepts in creating metrics.

Step 1. Selecting data

  1. On the Data and Metrics tab, select an existing project or create a new one by clicking the '+' button next to the project list.

    If you are creating a new project, remember to provide a unique name for the project.

  2. Click Create metrics.

  3. A new semantic data model is automatically created for you. Enter a unique name for the semantic data model. This is where you will model data later.

  4. Select the data that you want to work with.

    a. Create a new connection - Connect to your remote data by creating and testing a new connection.

    b. Use an existing data source - Connect to your remote data with an existing connection or choose an existing data asset such as an uploaded file.

    c. Use an IBM Cognos Analytics data source - Connect to your Cognos Analytics environment and import an FM package or Data module to work with.

  5. Click Next to continue to metadata enrichment.

    If you are using a Cognos Analytics data source, metadata enrichment is skipped. You are taken instead directly to Step 3: Creating metrics and data modeling.

Step 2: Metadata enrichment

  1. Click Run enrichment to start the metadata enrichment process.

    During metadata data enrichment, business context-based metadata is added to your data to prepare it for use in conversations and metrics.

    This metadata is generated by AI based on the predefined governance artifacts that are available in watsonx BI and any custom domain-specific glossary or business terms that you added.

    No changes are made to your original data during metadata enrichment.

  2. Review the results after metadata enrichment completes and make necessary changes to labels and descriptions, or accept the suggested ones.

    The enriched metadata results open in a separate tab and any changes you make save automatically.

    You don't need to re-run enrichment after making changes to the enrichment results.

    You must accept or edit the AI-generated labels and descriptions in order to use them during the metric creation process.

    Make sure that the labels and descriptions that are assigned to every column for all enriched data, are accurate and meaningful. AI uses labels and descriptions during conversations to find data to answer your questions. During metadata enrichment, AI-generated labels and descriptions might have also been assigned to your data.

  3. Close the tab after reviewing the results and return to the watsonx BI enrichment page and click Next to continue.

Step 3: Creating metrics and data modeling

  1. You are now on the Metrics overview page and can create metrics by:

    • Generating them automatically

      AI automatically generates metrics based on your enriched data. The generate metrics process does the following automatically for you:

      • creates metric definitions

      • enriches the newly-generated metric definitions

      • exports the selected metric definitions to the project as metrics for use in conversations

    • Building them manually

      You can build your own metrics in the semantic data model. To get started:

      a. Click a measure and select New > Metric definition.

      b. Enter details to define the metric.

      c. Save the semantic data model.

      d. Click the menu icon next to the newly-created metric definition and select Export metric definition.

      Exporting a metric definition creates a metric or updates an existing one in the project and makes it available for use in conversations.

      Anytime you make a change to the metric definition, for the change to apply to the metric in the project, you must export the metric definiton to the project.

  2. Navigate back to the Metrics overview page. You can see the metrics that you created and, if applicable, their enrichment status in the Metrics overview panel.

    You can now do the following:

    This is an iterative process. You can explore, adjust, and revisit these steps as needed.

    a. Try a metric in a conversation

    Click a metric and select Try metric to ask questions against the metric.

    b. Edit metrics

    You can modify a metric by editing its metric definition in the Advanced mode. If you tried a metric in a conversation and didn't get the expected responses, you can edit the metric definition and export it again for use.

    c. Use the Advanced mode to make changes to the semantic data model

    You can create new or modify existing relationships, calculations, metric definitions, and more. Learn how out to optimize data for AI in the semantic model.

    d. Add visualizations to metrics

    You can generate visualizations for a metric and choose the ones that you want to work with. You can also build visualizations manually.

    While visualizations are not mandatory, if you want a metric to display in a user's Key metrics, you need to create and associate a visualization to the metric. The visualization is what the user interacts with in the Key metrics panel.

    e. Publish metrics and visualizations to the Metrics catalog

    Select the metrics and associated visualizations that you want to publish to the Metrics catalog and specify access permissions.

    After the metrics are published to the Metrics catalog, Analytics consumers can start asking questions about this data.

If you want certain metrics to display in a user's Key metrics, you need to assign these to the user from the Metrics catalog.

Step 4: Assigning metrics to users

  1. In the Metrics catalog, select the metric that you want to assign.

  2. Select the visualization that will display in the Key metrics panel and click Assign to members.

  3. Select the users that you want to assign the metric to and click Save.

    To add new users and groups, click Add members.