The post demonstrates how to use Amazon SageMaker Feature Store Feature Processor to transform raw data into aggregated features suitable for training machine learning models, by automating feature engineering tasks like aggregation, encoding, and normalization at scale using Apache Spark.

<div>
<p>This article explains how to use the SageMaker Feature Store Feature Processor to engineer features from raw data for machine learning. It demonstrates how a car sales company can transform raw sales transaction data into aggregated features that provide valuable insights.</p>
<p>Specifically, the article covers:</p>
<ul>
<li>Creating feature groups for raw and aggregated car sales data</li>
<li>Using the @feature_processor decorator to load raw data into the raw feature group</li>
<li>Running the feature processing code remotely as a Spark application to aggregate the data</li>
<li>Operationalizing the feature processor via SageMaker pipelines and scheduling runs</li>
<li>Exploring the feature processing pipelines and lineage in SageMaker Studio</li>
<li>Conclusion on using Feature Processor to unlock ML insights from data</li>
</ul>
</div>


Unlock ML insights using the Amazon SageMaker Feature Store Feature Processor

Related articles