Get started with Data Explorer
Categories:
Prior reading: Data Explorer overview
Purpose: This document walks you through the basic features of Data Explorer, including how to create cohorts, build feature sets, and export your datasets.
Introduction
The goal of this guide is to show you the basic features of Data Explorer so that you can start working with datasets for detailed analyses.
Prerequisites
You'll need a Workbench account and access to an underlay in Data Explorer.
Log in to Data Explorer using your Workbench credentials.
If you see a “No underlays are configured” message, please reach out to your Workbench contact to gain access to the appropriate underlay.
In this guide, we'll use the cmssynpuf (Medicare Claims Synthetic Public Use Files) underlay.
Build a custom cohort
Once you’re logged in to Data Explorer, select your underlay's name. This will take you to a page that lists all of your studies for that particular underlay.
Select Add study. Enter a study name and select Create.
On the next screen, select the study you just created, and then select New cohort.
The main cohort page will show you two main sections: the Cohort filter and the Cohort visualizations.
Cohort filter
Use the Cohort filter to narrow down the criteria for your group. Select Add some criteria to display the various filter options.
You can either type in your criteria or scroll through various domain, program data, or source code fields.
To further refine your search, use the dropdowns to have your selected criteria included or excluded. You can also add multiple criteria and use the dropdown to meet any criteria (which triggers the “or” operator) or meet all criteria (which triggers the “and” operator).
In addition, you can select a temporal option from the dropdown. Temporal options only work with conditions criteria. If you’d like to filter your cohort by other criteria, such as ethnicity or age, create another group. Note that you can only use the AND operator when working with multiple groups.
Cohorts autosave, including when you rename the cohort and the groups.
In this example, we’ll narrow down the participants by selecting only those who are female, who have a low hemoglobin, and who have type 2 diabetes mellitus.

Cohort visualizations
The default cohort visualization view will be bar charts. You can hover over the bars to see a more detailed breakdown.

In the upper right of the Cohort visualizations section, you can turn certain charts on or off. In the upper right corner of each visualization, you can toggle between bar chart and table view.
These visualizations will update in real-time as you change the cohort's criteria.
Review participants
Once you’ve created your cohort, you can view individual participants and create annotations.
In the Cohort visualizations section, select Review individuals. This will take you to a Reviews page.
Select + New review and enter a name for your cohort review and the number of participants. Note that the maximum allowed is 9,999. Select Create. It will take a few moments to create the review. Once it's created, you'll see an Overview tab summarizing your cohort criteria and a Charts tab that shows the cohort visualizations.
To make notes about individual participants, you'll first need to set up annotation fields. Select the Annotations tab and then + Add annotation field to create preset fields. You'll see two preset fields:
- Free text allows you to enter any sort of annotations
- Review status will present a dropdown with Included, Excluded, and Needs further review options.
For all presets, you need to enter a display name for each annotation field. In this guide, we'll create one Free text preset named "Participant notes" and one Review status preset named "Review status."

Return to the Reviews tab and select Review individual participants. Here, you can view the selected participant’s conditions, procedures, observations, and drugs. On the left-hand side, you’ll see demographic information about the participant, as well as the annotation fields you created. You can navigate to different participants by selecting the arrow buttons, or you can view a specific participant by selecting the hamburger menu to open the full list of participants.
Any annotations you make will be saved automatically.

Select feature sets for exporting
Now that you’ve created a cohort with your desired criteria, you can build a feature set to export and further analyze your data.
Navigate to your study’s page and select New feature set. Feel free to give your feature set a custom name.
Select Add a data feature to add criteria to your feature set. In this example, we’ll select Procedure. You’ll see a list of various procedures; you can also view them in hierarchical form where values are nested. Learn more about OMOP data hierarchies and relationships here. In this example, we'll select Therapeutic procedure and then select Save criteria.
You'll then see a table with various procedures. By default, all table columns will be selected. You can deselect any columns that you don’t want to include in your feature set by either using the checkbox feature next to a column name or by selecting Manage columns and using the toggle function. The Show included columns only toggle can help provide a cleaner view of the columns you want to include in your feature set.

You can select as many other data features as you’d like. Feature sets will autosave.
Export your data
Once you’ve added your data features and the columns you want to include, you can export your data set.
Navigate to your study’s page and select Export. You'll see a page that lists all of your cohorts and feature sets for that particular study.
Select at least one cohort and one feature set. You'll see a summary of the dataset that will be exported. In this example, the summary shows the criteria we selected in our Diabetes cohort and the table we’ll get in the export based on the feature set we created. You can also select the Data tab to see your selected data either in table form or query form. If you want to update anything, you can select Edit next to your cohort or feature set. Otherwise, select Export dataset.

There are two export options: save to a Workbench workspace, or download a Jupyter notebook.
Save to a Workbench workspace
When you save to Workbench, you’ll be redirected to log in to Workbench and prompted to select the desired workspace and bucket to save the files. You can either use an existing bucket or create a new one.
On the following screen, you’ll see a list of files that will be saved to your workspace. You should see your cohort data as a CSV file and the feature set columns as a compressed CSV file. Select Save to Workbench to continue. Then you can select View in Workbench to see your exported data in your workspace.
Note
The compressed CSV file will have to be decompressed before it can be used for analysis.Download a Jupyter notebook
Alternatively, you can download your dataset as a Jupyter notebook. The file will download locally. In a running cloud app, you can drag and drop the downloaded file to JupyterLab and view it there.
Note that the downloaded notebook file includes the SQL queries, allowing you to play around with the query as needed. You can run the file in JupyterLab and it will create a Pandas Dataframe.
Last Modified: 2 June 2025