AWS Certified Data Analytics – Specialty — Question 163

An airline has been collecting metrics on flight activities for analytics. A recently completed proof of concept demonstrates how the company provides insights to data analysts to improve on-time departures. The proof of concept used objects in Amazon S3, which contained the metrics in .csv format, and used Amazon
Athena for querying the data. As the amount of data increases, the data analyst wants to optimize the storage solution to improve query performance.
Which options should the data analyst use to improve performance as the data lake grows? (Choose three.)

Answer options

• A. Add a randomized string to the beginning of the keys in S3 to get more throughput across partitions.
• B. Use an S3 bucket in the same account as Athena.
• C. Compress the objects to reduce the data transfer I/O.
• D. Use an S3 bucket in the same Region as Athena.
• E. Preprocess the .csv data to JSON to reduce I/O by fetching only the document keys needed by the query.
• F. Preprocess the .csv data to Apache Parquet to reduce I/O by fetching only the data blocks needed for predicates.

Correct answer: C, D, F

Explanation

The correct answers are C, D, and F. Compressing the objects reduces the amount of data transferred, improving I/O performance. Using an S3 bucket in the same Region as Athena minimizes latency and enhances query speed. Converting .csv to Apache Parquet optimizes storage and allows for more efficient data access by fetching only necessary data blocks. Options A and E do not provide significant performance improvements compared to the selected options.