This directory has notebooks containing samples and explanations of common workflow patterns for updating datasets in Amazon Personalize.
Both samples notebooks contain an end-to-end workflow pattern including required setup and cleanup. These notebooks were developed and tested on an Amazon SageMaker Notebook Instance.
- Launch a SageMaker Notebook Instance
- Ensure the Execution Role of your Notebook instance has the required Amazon IAM permissions. A sample policy document that grants appropriately-scoped permissions is defined in sagemaker_notebook_exec_role.json. You can attach this custom policy as a customer-managed policy of your Execution Role.
- Upload the notebook to the SageMaker Notebook Instance
- Run the Notebook from the Jupyter environment
Notebook update-item-dataset-schema-example.ipynb goes over the process for updating schemas of your datasets in Amazon Personalize; specifically schemas for items. Commentary in this notebook is centered around the an e-commerce use case that uses the item-attribute-affinity recipe.
Notebook update-datasets-user-personalization-example.ipynb goes over the process for updating your datasets when using Amazon Personalize; specifically adding new items and interactions for those new items. Commentary in this notebook is centered around the an e-commerce use case that uses the user-personalization recipe. Thus, the auto-update feature for user-personalization solution versions is also discussed.
This sample code is made available under a modified MIT license. See the LICENSE file.