Although there are 45 stores, we will select one store, store number 20, to analyze performance across different departments across three years. The main idea here is that, using DeepAR, we can learn the sales of items across different departments.
In SageMaker, through Lifecycle Configurations, we can custom install Python packages before notebook instances are started. This eliminates the need to manually track packages required before the notebooks are executed.
For exploring the retail sales data, we will need the latest version, 0.9.0, of seaborn installed.
In SageMaker, under Notebook, click on Lifecycle Configurations:
- Under Start notebook, enter the command to upgrade the seaborn Python package, as shown: ...
