You work as a machine learning specialist for a mobile phone operator where you need to build a machine learning model that predicts when a given customer is about to leave your phone service or churn. The inference data produced by your model will allow your marketing department to offer incentives to the customer to get them to stay with your service. Using data generated by customer activity with your service offering, you need to visualize the inference data in a dashboard. So your marketing department can quickly decide which customer churn candidates to offer additional incentives. How can you get your machine learning inference data into your dashboard visualization in the most efficient, performant manner?

Answer options:

A.As your inference engine produces potential churn candidate data, write the data to S3. Use Athena to query the data and associate a QuickSight visualization data source with your Athena query results.
B.Create a JSON schema file that contains the metadata that QuickSight needs to process your model data, then use the Augment with SageMaker feature of QuickSight to visualize your customer churn data.
C.As your inference engine produces potential churn candidate data, write the data to S3. Use S3 Analytics to query the data and associate a QuickSight visualization data source with your S3 Analytics query results.
D.As your inference engine produces potential churn candidate data, write the data to S3. Use Redshift Spectrum to query the data and associate a QuickSight visualization data source with your Redshift Spectrum query results.

Correct Answer: B
Option A is incorrect. Although you could use Athena and S3 to visualize your data as this option describes, it is not as efficient as using the Augment with SageMaker feature of QuickSight.
Option B is correct. You can use the Augment with SageMaker feature of QuickSight to integrate your SageMaker inference data into your QuickSight visualization. This is by far the most efficient option given.
Option C is incorrect. Although you could use S3 Analytics and S3 to visualize your data, as this option describes, it is not as efficient as using the Augment with SageMaker feature of QuickSight.
Option D is incorrect. Although you could use Redshift Spectrum and S3 to visualize your data, as this option describes, it is not as efficient as using the Augment with SageMaker feature of QuickSight. You would have to create a Redshift cluster for this option to work.
References:
Please see the AWS Machine Learning blog titled Visualizing Amazon SageMaker machine learning predictions with Amazon QuickSight (https://aws.amazon.com/blogs/machine-learning/making-machine-learning-predictions-in-amazon-quicksight-and-amazon-sagemaker/),
The AWS Samples GitHub repo titled ML Predictions using Amazon QuickSight and Amazon SageMaker (https://github.com/aws-samples/quicksight-sagemaker-integration-blog)