Creating a data profile

Applies to: Kyvos Enterprise  Kyvos Cloud (SaaS on AWS) Kyvos AWS Marketplace

Kyvos Azure Marketplace   Kyvos GCP Marketplace Kyvos Single Node Installation (Kyvos SNI)

A data profile is a sample analysis of data that you can use to design a semantic model, or dataset that's more responsive to queries. The data profile also helps recommend which advanced properties and process options you may want to use. Not all object types support a data profile.

• By default, it runs on the column level looking for distinct value counts, value distribution (is the data skewed), cardinality, and seeks invalid value counts.

• You can view details about the data, including field names, field types, distinct value counts, and value distribution. For numeric fields, you'll see the minimum and maximum values and the number of invalid values.

• You can also scroll to a column, view a summary, download data profile logs, or export the profile result.

• Additionally, you can change the default profile selection details.

To create a data profile, perform the following steps. 

1. From the Toolbox area, select one of the following:

   1. Semantic models: For semantic models, select a semantic model in the Process tab, click the Actions menu (...), and then click Add Other Jobs.
      
      

      Open

      

   2. Files: For files, click the ... menu and click Add Job.

2. Choose Data Profile.

3. If desired, you can change the default profile selection.

4. Choose which records to use: all records for complete analysis or use a random sample.

5. Schedule when to run the profile.
   Typically, you will run a profile only a few times, so you may want to start with Now or Once.

6. Click Schedule when finished.

   Open

   

Viewing profile results

Depending on the object type, you may see slightly different options.

To view semantic model profiles from the semantic model, click the Process tab. From the Actions menu (...) choose Data Profile Result.

To view file profiles, do one of the following:

• Right-click a file and click Last Profile Result.

• From the Actions menu (...) menu click Last Profile Result.

• From the Add Job dialog box, click History, right-click a Data Profile job in the list, and choose View Profile Summary.

When viewing the profile, you can:

• Click Scroll to Column and scroll or use Search to find a specific value.

• Click Summary to view the summary of the process. Then click Close.

• Click Download Logs to download the process logs.

• Click Export to export the data to your local machine.

Related topics

• Customizing a semantic model profile selection

• Using semantic properties