Sagemaker tutorial - Prepare Training Data for Machine Learning with Minimal Code - fails to deploy

The error you're encountering is due to the KernelGateway app type being deprecated as part of AWS's transition from SageMaker Studio Classic to the new SageMaker Studio.

Based on the available information, there isn't a direct replacement app type specified for KernelGateway in the newer SageMaker Studio. This is part of the architectural changes between SageMaker Studio Classic and the new SageMaker Studio.

Since this tutorial appears to be using older CloudFormation templates designed for SageMaker Studio Classic, you have a few options:

1. Check if AWS has released an updated version of this tutorial that works with the new SageMaker Studio

2. Consider using a different approach to set up your environment:

• You could manually create a SageMaker domain and user profile through the AWS console
• Then launch SageMaker Studio and create a notebook instance there

3. Contact AWS Support for specific guidance on migrating this tutorial to the new SageMaker Studio architecture

If you need to work with SageMaker immediately, you might want to explore other tutorials that have been updated for the current SageMaker Studio environment, such as the end-to-end machine learning workflow tutorial in SageMaker Canvas, which provides a visual no-code interface for preparing data and training ML models.
Sources
Troubleshooting Amazon SageMaker Studio Classic - Amazon SageMaker AI
Tutorial: Build an end-to-end machine learning workflow in SageMaker Canvas - Amazon SageMaker AI

You're absolutely right about the KernelGateway app type being deprecated as part of the transition from SageMaker Studio Classic to the new SageMaker Studio experience.

Here's how to fix your template:

Replace the DataScienceApp and DataWranglerApp resources with:

DataScienceApp:
    Type: AWS::SageMaker::Space
    DependsOn: StudioUserProfile
    Properties:
      DomainId: !GetAtt StudioDomain.DomainId
      OwnershipSettings: 
        OwnerUserProfileName: !Ref UserProfileName
      SpaceName: instance-event-engine-datascience-ml-t3-medium
      SpaceSettings: 
        AppType: JupyterLab
        SpaceStorageSettings:
          EbsStorageSettings:
            EbsVolumeSizeInGb: 50
        JupyterLabAppSettings:
          DefaultResourceSpec:
            InstanceType: ml.t3.medium

DataWranglerApp:
    Type: AWS::SageMaker::Space
    DependsOn: StudioUserProfile
    Properties:
      DomainId: !GetAtt StudioDomain.DomainId
      OwnershipSettings: 
        OwnerUserProfileName: !Ref UserProfileName
      SpaceName: instance-event-engine-datawrangler-ml-m5-4xlarge
      SpaceSettings: 
        AppType: JupyterLab
        SpaceStorageSettings:
          EbsStorageSettings:
            EbsVolumeSizeInGb: 50
        JupyterLabAppSettings:
          DefaultResourceSpec:
            InstanceType: ml.m5.4xlarge

This update replaces the deprecated KernelGateway app type with the new Space resource type, which is the recommended approach in the new SageMaker Studio.