Main goal is to give students a sense of why they might need cloud services and, when they do, what their options are.

Learning goals listed at the top:

• Explain the general concept of "the cloud";
• Understand the cases where hardware acceleration is useful;
• Understand the cases where cloud storage and the Boto3 library in particular are useful;
• Explain the purpose of deploying a machine learning model, particularly with the Flask library.

Summary at the end of the main lecture content:

• Cloud services are useful for computationally intensive or long-running tasks
• The major providers of cloud services are Amazon Web Services (AWS), Microsoft Azure, and Google Cloud
• As a data scientist, you will generally use cloud services to get more computing power and/or to deploy machine learning models
• If you want to get more computing power, consider:
  • Cloud instances/containers with GPUs, particularly EC2
  • Cloud notebooks, particularly Google Colab
  • Cloud storage, particularly S3 bucket storage with Boto3
• If you want to deploy a machine learning model, first pickle the model, then consider:
  • Deploying a model as an API, using either a cloud function or a minimal Flask app
  • Deploying a model as a full-stack web app, either using Flask or Dash

Jupyter Notebook: Cloud Services