Question #1451
An ecommerce company aims to deploy machine learning (ML) models to predict customer churn and integrate these predictions into their business intelligence (BI) dashboards. The architecture team seeks a solution that minimizes operational overhead by leveraging fully managed AWS services for both model training and visualization. Which approach will MOST effectively meet these requirements?
Use Amazon Kinesis Data Analytics for real-time model training and Amazon OpenSearch Service for dashboard visualization.
Use Amazon SageMaker to build and train models, then visualize insights using Amazon QuickSight.
Train models using a custom ML framework on Amazon EC2 instances and visualize data with Amazon Managed Grafana.
Implement AWS Glue ML Transforms for model training and use Amazon Redshift dashboards for visualization.
Explanation
Answer B is correct because:
1. Amazon SageMaker is a fully managed service for ML model development, training, and deployment, eliminating the need for infrastructure management.
2. Amazon QuickSight is a serverless BI tool that integrates seamlessly with AWS data sources, enabling easy visualization of predictions.
Other options are incorrect because:
- A: Kinesis Data Analytics focuses on real-time streaming analytics, not ML model training. OpenSearch is better suited for search/analytics than BI dashboards.
- C: EC2 requires manual infrastructure management, increasing operational overhead. Managed Grafana is less tightly integrated with AWS analytics services.
- D: AWS Glue ML Transforms are limited to basic transforms (e.g., normalization), not full ML model training. Redshift dashboards lack QuickSight's advanced BI features.
Key Points:
- Use SageMaker for end-to-end ML workflows.
- QuickSight is AWS's native BI service for dashboards.
- Avoid non-managed services (e.g., EC2) to reduce operational effort.
Answer
The correct answer is: B