Overview of applications in the Researcher Workbench

Recognizing that researchers want to analyze data in different ways depending on their specific needs, the Researcher Workbench supports a variety of cloud-based applications to analyze All of Us participant data. Unlike the Cohort and Dataset Builders, applications incur costs because they require cloud environments to use. A cloud environment refers to a virtualized infrastructure that allows users to access and manage computing resource over the internet. These applications can be used to analyze participant data that's generated manually or with the Cohort and Dataset Builders.

Below is an up-to-date list of the different analysis tools along with information about associated cloud environments, basic demonstration videos, support materials, and other relevant information. This article will be updated as other applications are made available in the Workbench.

• Jupyter Notebook
• RStudio
• SAS Studio
• Cromwell

Jupyter Notebook

Overview

Jupyter Notebook is a popular tool used by data scientists, researchers, and developers for interactive computing and data analysis. They provide an environment where users can write and execute code, visualize data, and document their work in a single document. R or Python notebooks are supported in the Researcher Workbench, allowing users to choose the language that best suits their needs. These notebooks consist of a series of cells, each of which can contain code, text, equations, or visualizations. Users can run individual cells or the entire notebook, making it easy to experiment with code and see the results in real-time. 

Environments

Jupyter Notebook are powered by Jupyter environments, which can be enabled in the Jupyter icon in the right panel of a workspace as shown below. 

A variety of customization options are available, depending on your computing needs. If you'd like to understand more about customizing and deleting your Jupyter environment, including about using Dataproc clusters for Hail analyses, please see this article that details different aspects of cloud environments and how to optimize them.

Environments can be created by clicking the create environment button in the bottom right corner.

In addition to Jupyter Notebook, Jupyter environments can be used when analyzing data using the terminal.

Support resources

Since Jupyter Notebook have been supported since launch, most of our support materials are designed around their use. This includes our featured workspaces (requires Researcher Workbench login), which are nearly all designed to use Jupyter Notebook. Here are a few resources that are specifically related to getting started in Jupyter Notebook.

• An article about analyzing data in Jupyter Notebook.
• Customizing your Jupyter environment.
• Costs associated with running Jupyter environments.
• Storage options for Jupyter environments.
• An introductory featured workspace showing how to analyze data (requires Researcher Workbench login).

RStudio

Overview

RStudio is an integrated development environment (IDE) for R, a programming language used for statistical computing and graphics. It provides a user-friendly interface that makes it easier for users to write, debug, and execute R code. RStudio offers a wide range of features and tools, including a code editor with syntax highlighting and auto-completion, a console for executing R commands, a workspace for managing objects and datasets, and a plotting window for visualizing data. It also supports the creation of interactive documents and reports through its integration with R Markdown. 

Environments

RStudio is used in conjunction with RStudio environments, which can be created by clicking the R icon on the right hand side of a workspace as shown below. Environments can't be customized and cost a fixed $0.40 when running.

Support resources

• A featured workspace notebook that demonstrates how to use RStudio (requires login to the Researcher Workbench)
• Support articles about using RStudio in the Workbench and how to export datasets into RStudio.

SAS Studio

Overview

SAS is a statistical software tool used for data analysis and statistical modeling. The SAS Studio application in the Researcher Workbench connects to a SAS server in order to process SAS commands. The SAS server is hosted in a cloud environment. After code is processed by the SAS server, the results are returned to the SAS app in your workspace. SAS has the ability to handle large datasets efficiently, with an extensive set of statistical techniques, and a user-friendly interface. 

Environments

SAS Studio on the Researcher Workbench uses a cloud environment, which may be slightly different than using a SAS Studio application on your local computer. SAS Studio runs on a virtual machine (VM) or clusters of machines in your workspace cloud analysis environment. The SAS cloud environment on the Researcher Workbench is a VM with 4 CPUs, 15 GB RAM, and 250 GB of disk space and is not customizable. 

To start SAS Studio, follow the instructions outlined in this support article. 

Support Resources

We offer several support resources to help you get started using SAS in the Researcher Workbench such as: 

• How to run SAS in the Researcher Workbench
• Exploring All of Us data using SAS Studio
• Using the Cohort Builder & Dataset Builder with SAS
• Analytics Guides
  • • SAS Analytics Guide - How to perform logistic regression
    • SAS Analytics Guide - How to perform binary logistic regression
    • SAS Analytics Guide - How to Estimate Frequency

Cromwell for workflows

Overview

Cromwell is a workflow management system that is designed to help scientists and researchers organize and execute complex computational workflows. It provides a platform for defining, running, and monitoring workflows, making it easier to manage and automate scientific analyses. With Cromwell, users can define their workflows using a simple and intuitive syntax, allowing them to specify the steps and dependencies of their analysis. This makes it easier to break down complex tasks into smaller, more manageable units, improving reproducibility and scalability. Cromwell also provides a range of features to enhance workflow execution. It supports a variety of execution backends, allowing users to run their workflows on different computing infrastructures, such as local machines, clusters, or cloud platforms. It provides detailed logs and metrics, allowing users to troubleshoot issues and optimize their workflows for better performance. 

Cromwell environments

Workflows can be submitted in the Researcher Workbench through Cromshell, which is a command line tool for interacting with Cromwell. They can be created through the flying pink pig Cromwell icon on the right side of the workspace panel as described in this support article. You can then start an environment by clicking the Start button.

Cromwell environments aren't customizable on the Workbench, though using Workflow Description Language (WDL) you can customize the way a workflow is run. In order to use Cromshell, you'll need to do so through a Jupyter Notebook or via the terminal, which require a Jupyter environment to power. 

Support resources

• Below is a video that walks through working with Cromwell in the Researcher Workbench.

• A featured workspace that demonstrates how to use Cromwell (requires login to the Researcher Workbench)
• An article about Cromwell and how to submit workflows.