MLA-C01 Question 84
Select 2You are building a machine learning model to predict customer churn for a subscription-based service. The dataset contains several numerical and categorical features, but during an initial analysis, you notice missing values, duplicate entries, and inconsistent data formats. How should you ensure data integrity and prepare the dataset for modeling using AWS services?
- A
Use AWS Glue to clean and normalize the data, ensuring consistent formats and removing duplicates.
- B
Use Amazon Athena to query the dataset and identify missing values, then manually fill them in the source file.
- C
Apply Amazon SageMaker Data Wrangler to detect and handle missing values, duplicates, and data type inconsistencies.
- D
Use AWS Batch to automate the dataset cleaning process by running custom Python scripts.
- E
Leverage Amazon QuickSight to visualize the dataset and identify anomalies, then remove the anomalies directly within QuickSight.
Show answer and explanation
Correct answers: A, C
Explanation
Ensuring data integrity and preparing data for modeling requires leveraging the right AWS services for cleaning and preprocessing. AWS Glue is a powerful ETL service for data cleaning, and SageMaker Data Wrangler is a specialized tool for preparing datasets for machine learning workflows. These services handle common issues like missing values, duplicates, and inconsistent formats efficiently. Other options like Athena and QuickSight are valuable for analysis but not designed for direct data cleaning, while AWS Batch is not tailored for this purpose.
- A. Correct.
AWS Glue is a fully managed ETL (Extract, Transform, Load) service that can be used to clean and normalize datasets. It is well-suited for removing duplicates and standardizing data formats, ensuring data integrity.
- B. Incorrect.
While Amazon Athena can query the dataset and identify missing values, it is not designed for manual edits or filling missing values directly in the source file. This is not an efficient or scalable approach.
- C. Correct.
Amazon SageMaker Data Wrangler is specifically designed to help clean, transform, and prepare datasets for machine learning. It provides built-in functionalities to handle missing values, duplicates, and inconsistencies in data types.
- D. Incorrect.
AWS Batch is a service for running batch computing workloads, but it is not specifically designed for data cleaning. Using custom Python scripts with AWS Batch may work, but it is not the best or most efficient solution for this scenario.
- E. Incorrect.
Amazon QuickSight is a visualization tool and does not provide the capability to directly clean or modify datasets. It is useful for identifying anomalies but not for handling them directly.