Offload infrequently used data and batch load raw data to a defined landing zone in an Amazon S3 bucket. For a data lake, you load raw transactional and multi-structured data directly to Amazon S3, freeing up space in the current Enterprise Data Warehouse..

Collect and stream data generated by machines and sensors, including application and weblog files, directly to Amazon S3 instead of staging it in a temporary file system or the data warehouse.

Discover and profile data stored on Amazon S3. Profile data to better understand its structure and context. Adding requirements for enterprise accountability, control, and governance for compliance with corporate and governmental regulations and business service level agreements.

Parse and prepare data from weblogs, application server logs or sensor data. Typically, these data types are either in multi-structured or unstructured formats which can be parsed to extract features and entities, and data quality techniques can be applied. You can execute pre-built transformations and data quality and matching rules natively in Amazon EMR to prepare data for analysis.

After you cleanse and transform data on Amazon EMR, move high-value curated data from EMR to an Amazon S3 output bucket or to Amazon Redshift. From there, users can directly access data with BI reports and applications.