The correct choice is Amazon SageMaker Data Wrangler. It provides a visual end to end workflow to import join cleanse and transform data from many sources and to prepare user profile inventory and clickstream datasets for ingestion into Amazon Personalize.

Amazon SageMaker Data Wrangler includes built in connectors and transformation utilities so teams can unify and clean records coming from twelve different systems and then export the prepared data to Amazon S3 or to downstream ML services for training and recommendations.

Amazon SageMaker Clarify focuses on bias detection and model explainability and it does not perform broad data ingestion and transformation across heterogeneous sources.

Amazon SageMaker Ground Truth is designed for data labeling and human in the loop annotation and it is not intended for joining cleaning and transforming large tabular datasets for recommendation pipelines.

Amazon SageMaker Feature Store is meant to store and serve ML features for training and inference and it does not handle the upstream ingestion and cleaning steps required to unify raw source data for Amazon Personalize.

Full AWS Practitioner Certification Question