Amazon SageMaker Feature Store

Introduction to Feature Store

In feature_store_introduction.ipynb we demonstrate how to get started with Feature Store, create feature groups, and ingest data into them.

This notebook requires these data sets in ./data/:

• feature_store_introduction_customer.csv
• feature_store_introduction_orders.csv

This notebook requires these images in ./images/:

• feature-store-policy.png
• feature_store_data_ingest.svg.

Feature Store Encryption with KMS key

In feature_store_kms_encryption.ipynb we demonstrate how to encrypt data in your online or offline store using KMS key and how to verify that your KMS key is being used for data encryption.

This notebook requires these data sets in ./data/:

• feature_store_introduction_customer.csv
• feature_store_introduction_orders.csv

This notebook requires these images in ./images/:

• cloud-trails.png
• s3-sse-enabled.png

Client-side Encryption using AWS Encryption SDK

In feature_store_client_side_encryption.ipynb we demonstrate how client-side encryption with SageMaker Feature Store is done using the AWS Encryption SDK library to encrypt your data prior to ingesting it into your Online or Offline Feature Store. We first demonstrate how to encrypt your data using the AWS Encryption SDK library, and then show how to use Amazon Athena to query for a subset of encrypted columns of features for model training.

This notebook requires this synthetic data set in ./data/:

• credit_card_approval_synthetic.csv

Securely store an image dataset in your Feature Store with KMS key

In feature_store_securely_store_images.ipynb we demonstrate how to securely store a dataset of images into your Feature Store using KMS key.

Securely store the output of an image or text classification labelling job from Amazon Ground Truth directly into Feature Store using a KMS key

In feature_store_classification_job_to_ground_truth.ipynb, we demonstrate how to pipe the output of an image or text classification labelling job from Amazon Ground Truth directly into Feature Store.

Fraud Detection with Feature Store

For an advanced example on how to use Feature Store for a Fraud Detection use-case, see Fraud Detection with Feature Store, and it's associated notebook, sagemaker_featurestore_fraud_detection_python_sdk.ipynb.

Developer Guide

For detailed information about Feature Store, see the Feature Store Developer Guide.