Getting Started with All of Us Data

Programming Tutorials

All of Us Researcher Workbench

1. Registration and Training

• Go to the All of Us Research Hub. Click on the “REGISTER” button on the top right corner and follow the instructions to verify identity and create an account.
• Complete required training modules. University of Idaho has Controlled Tier level access. You must complete the second (longer) training module to gain access to the Controlled Tier data.

2. Learn about Data

• Watch New User Orientation recording:
• Read support articles on Understanding the All of Us Data:
  • Data types and organization
  • Participant privacy protections
  • Understanding OMOP basics
  • Exploring concepts with OMOP and SQL (You can gnore SQL section unless you will be using manual SQL queries)

3. Learn to use Researcher Workbench

• Create a Workspace for your project
  • Step-by-step Creating a Workspace
    • NOTE When selecting data access tier, make sure to choose Controlled Tier. In most cases, this is what we want. For Dataset Version, choose v8 (whatever the latest)
  • Writing Your Workspace Description
  • Sample Workspace Description
• Learn to use Cohort Builder and Dataset Builder
• Familiarize with Jupyter Notebook
  • Researcher Workbench offers several apps: Jupyter Notebook, R Studio and SAS Studio. However, I recommend using Jupyter Notebook due to less cost and better stability.
  • Jupyter Notebook in 10 Minutes (Don’t worry about installing and running it. Just watch and get acquainted with the interface)
• Some notes on storage options:
  • The Workbench offers two storage options: Persistent Disk and Workspace bucket.
  • Persistent disk is like a thumb drive: it attacheds automatically every time you spin up an environment. However, I discourage you from relying on the persistent disk for following reasons:
    1. Additional cost (~$4 per month)
    2. Collaborators cannot access your files
  • A “bucket” is a permanent storage space on Google Cloud Platform associated with each Workspace. Accessing the bucket requires gsutil commands. It only requires 1-2 lines of code and I will have R or Python template codes in the Workspace for you.
  • Best Practice: Write any cleaned/important/final data frames to a CSV file and then save it in the bucket. When finished, remember to check “Delete persistent disk” option when deleting the environemnt.
  • For more information on storage, check out this support page

4. More about EHR

5. More about Survey Data

Additional Resources