The correct choice is Import the data into Amazon SageMaker Canvas and create a demand prediction model using its point-and-click workflow. This option meets the requirement for a simple no-code method that lets business users combine internal sales history and third party market data and compare predictive models for weekly demand across multiple locations.

Amazon SageMaker Canvas is purpose built for non ML experts and it provides a visual, point and click workflow to prepare tabular datasets build and evaluate forecasting models and generate predictions without writing code. Canvas can ingest data from common sources and it produces results that analysts can review or export for downstream use.

Store data in Amazon S3 and train a forecasting model with Amazon SageMaker built-in algorithms is not ideal because that approach requires setting up training jobs and deployment with SDKs or notebooks and it does not satisfy the no code constraint.

Prepare datasets in Amazon SageMaker Data Wrangler and build models using SageMaker built-in algorithms is unsuitable because Data Wrangler focuses on data preparation and you still need code driven training to run built in algorithms which breaks the requirement for a no code workflow.

Import the data into Amazon SageMaker Data Wrangler and try to generate demand predictions with the Amazon Personalize Trending-Now recipe is incorrect because Amazon Personalize is a recommender service and the Trending Now recipe targets recommendations rather than general time series forecasting or demand prediction.

Full AWS Practitioner Certification Question