Creating a Data Model (CX)

About Creating a Data Model

The data modeler can be used to create datasets where you flexibly combine data from multiple types of sources. The data modeler has nearly all of the features the data mapper does, but with some additional advanced functionality, such as left outer data joins. The data modeler also makes it easier than ever to map not just survey data, but data from your directory, tickets, and other data sources, while allowing you to combine them in the same dataset.

Incompatible Features

The following features aren’t compatible with data models yet:

• Page templates
• Date time segmentation
• Spotlight Insights
• Org hierarchies
• XM Respondent Funnel page templates (XM Respondent Funnel dataset is compatible)
• Distribution reporting

Creating a Dataset with the Data Modeler

1. Create a new dataset.
2. Under Type, select Data modeler.
   
3. Name your dataset.
4. Click Create dataset.
5. Click Add Source.
   
6. Under Source Data, choose a survey, ticket set, directory, or another external data source. Use the search bar and All source types drop-down to find a source.
   
   Qtip: Only tickets created from surveys can be mapped with the data modeler. You need to be either the owner of the associated survey or a collaborator with both Edit and View reports access.
   Attention: Only projects you have access to will appear in this list. If any projects aren’t showing up, make sure you have been invited to collaborate on them.
   Attention: Legacy distribution reporting is not compatible with the data modeler.
7. When you’re ready to select a source, adjust the Quantity to 1 (or more, if desired).
   Qtip: It can be useful to add the same source more than once if you plan to perform separate unions and joins in your dataset. You cannot create a union or a join with multiple copies of the same source.
8. Select all the fields you want to include in your dataset. You can click on different types of data, like “Metadata”, to see only those fields, or use the search bar to search for specific fields.
   Qtip: When selecting fields to include in your data model, think about which fields will be important when analyzing your data. Fields that are commonly included are survey questions, important metadata, and the field you plan on using to join your datasets.
9. Add more data sources, or perform joins and unions. See the linked pages for more details.
   
   Qtip: If you want to report on 2 different surveys with 2 years worth of feedback data, you’ll most likely use a union. If you want to create a list of all of the customers who reached out across those years, you’d use a join.
10. Once you’ve finished, click the plus sign ( + ).
11. Select Output Data Set.
12. Check the Dataset Editor for accuracy. Here, you can find all of your data fields.
    
    Qtip: If needed, you can always edit these fields later.
    Qtip: Click Close to finish previewing your output dataset.
13. Select Publish.
    
    Qtip: Publishing will take you back to the Data page. See how to add your dataset to a dashboard.

Types of Sources Available in Data Models

A dataset gets its data from specific sources in Qualtrics. Data sources are often surveys, but there are many other sources you can add to data models. The following source types are available:

• DX Sources: Sessions from Digital Experience Analytics.
• Directories: XM Directory contacts.
  Qtip: Directory data requires additional setup before it can be used in your data models. See Enabling the Contact Dataset.
• External Datasources: Everything that doesn’t fit into the other categories listed here, such as imported data projects, segment membership data, and some older reputation management projects.
  Qtip: Segment data requires additional setup before it can be used in your data models. See Using Segment Data in Dashboards.
• Locations: Location directories.
• Reviews: Reputation Management programs.
• Surveys: Survey projects.
• Tickets: Tickets created from a tickets task in a survey. This does not include all tickets in the brand.
  Qtip: Only tickets created from surveys can be mapped with the data modeler. You need to be either the owner of the associated survey or a collaborator with Edit or View reports.

Continuous vs. Periodic Dataset Updates

For the most part, data model datasets update continuously. That means that as you collect more data in your sources, your dashboard receives that data accordingly, including all unions and joins and field remappings. Continuous updates ensure your dashboard’s data stays current.

However, certain configurations in your data model can make it update periodically instead. Periodic updates exist to lighten the load on your dataset and ensure accuracy when particularly complex setups are involved. However, to get everything correct, the updates are slower.

Periodic datasets generally update with new data weekly. Periodic datasets cannot receive new data before the weekly update, even after recaching. This schedule cannot be customized.

Identifying Features that Need Periodic Updates

Here are a few examples of features that cause a dataset to become periodic instead of continuous:

• Inner joins
• Outer joins
• Aggregations
• Chained joins

This list is not exhaustive. If your dataset updates periodically instead of continuously, there are a few places you can look in the data modeler to find this out:

Dashboards You Can Add Data Models To

Data models can only be added to the following types of dashboards:

• Dashboards projects (also known as CX Dashboards)
• Dashboards added to programs
• Dashboards built into certain XM Solutions
• Brand Experience

See Dashboard Data for steps to add your completed data model to your dashboard.

This feature cannot be used with any Employee Experience dashboards or with Results Dashboards. For a similar Employee Experience feature, see Data Models.

FAQs