Amazon Web Services (AWS), this week announced the launch Glue DataBrew. This tool allows organizations to prepare their data for machine-learning projects using a simple point and click interface.
Glue DataBrew, AWS’ original Glue product that was introduced in 2017, is an extension to this product. Glue was originally created to automate the tasks of extracting, transform and loading (ETL) data in preparation for machine learning projects. AWS estimates that this process can take several months. Glue promised that it would be completed in minutes.
AWS describes Glue DataBrew as a “visual data preparation instrument” that simplifies ETL by helping organizations “cleanse and normalize data upto 80% faster” through its visual interface.
This tool can access data from multiple AWS resources including Amazon Redshift, Amazon Relational Database Service (S3), and the Glue metadata storage. It can also access data stored in Java Database Connectivity API.
The Glue DataBrew console can be used by organizations to organize, combine, and manage their data. There are 250 “transformations” available that automate common data cleanup tasks, such as detecting anomalies, correcting nonstandard formatting, or removing invalid characters.
Users can apply any transformations they want to future data automatically. AWS announced that each transformation is added as a step in building a recipe. “You can save, publish, and modify recipes and automate data preparation tasks by applying recipes to all incoming data.
Glue DataBrew also offers a “lineage map” that visually tracks the transformations applied to a set of data. “This allows you to see how data flows and what the changes are.” AWS stated. “This information is known as data lineage and can be used to help you identify the root cause of errors in your output.”
Glue DataBrew can be found in the following AWS regions. According to AWS, users pay for the service as they use it. There is no upfront commitment. You can find more information about the service here.
