In this chapter, we leveraged the power of the AWS ecosystem to build a real-time streaming data classification pipeline. Our pipeline was able to classify streaming tweets using an Amazon ML classification endpoint. The AWS data ecosystem is diverse and complex, and for a given problem, there are often several possible solutions and architectures. The Kinesis-Lambda-Redshift-Machine Learning architecture we built is simple, yet very powerful.
The true strength of the Amazon ML service lies in its ease of use and simplicity. Training and evaluating a model from scratch can be done in a few minutes with a few clicks, and it can result in very good performances. Using the AWS CLI and the SDK, more complex data flows and model explorations ...