In this chapter, we focused on what is commonly known as the Extract Load Transform (ETL) part of the data science flow with regard to the Amazon ML service. We saw that the Amazon ML datasource is a set of information comprised of location, data structure, and data analytics given to the service so that it can use that data to start training models. You should now feel comfortable creating an Amazon ML datasource from an original CSV data file made accessible via S3.
We have also explored ways to transform the data and create new features via the AWS Athena service using simple SQL queries. The ability to complement the features of Amazon ML by leveraging the AWS ecosystem is one of the main benefits of using Amazon ML.
We now have a couple of Titanic datasets, the original one and the extended one, which are split into training and held-out subsets, and we have created the associated datasources.
In Chapter 5, Model Creation, we will use these datasets to train models, and we will...