Jupyter Workspaces

Intro

Jupyter Workspaces in Domo is a web-based interactive development environment for Jupyter notebooks, code, and data. Workspaces are tightly integrated with Domo to allow users to easily explore their Domo DataSets, leverage instantaneous code execution to develop pipelines for data science and machine learning, document their processes, create custom visualizations, and write transformed data back into Domo.

Jupyter Workspaces use the JupyterLab interface. For more information on JupyterLab, see JupyterLab Documentation.

Enabling Jupyter 

In order to start using Jupyter, you must enable it by completing the following.

1. Select Admin from the navigation header. For smaller screen sizes, you may need to expand the More tab. 

 

2. Navigate to Features and select Jupyter from the options. 

 

3. Activate the feature by adjusting the toggle next to Jupyter Account Inactive.

Your account has been activated.

After activating the account, choose the account plan you want:

The Default account plan allows all users access to Jupyter and gives all users unlimited usage.
The Manual account plan allows specific users access to Jupyter and gives you the option to configure usage limits.

4. Select a Compute Tier Limit.

5. Select Save.

 

Creating a Workspace

1. After enabling Jupyter by following the steps provided earlier in this article, access the feature by selecting Data in the navigation header.

 

2. In the left side rail, select  More and choose Jupyter Workspaces.

 

3. Create a new workspace by selecting + New Workspace.

The Create Jupyter Workspace modal displays. 

 

4. Customize the workspace by adding a Title, Description, Compute Tier Limit, Timeout, Kernel, and Input/Output DataSets. 

 

Learn more about Workspace Options in the table below: 

 

Title	The title of the Jupyter workspace
Description	An optional description to provide more details about the workspace.
Compute Tier Limit	The computer size that will be allocated to the workspace and any data flows that are associated with this workspace.
Timeout	The amount of time with no user activity in the Jupyter UI before the workspace will automatically stop.
Kernel	Python or R Kernels are available
Start workspace on successful creation	Once created, the process to start the workspace will be performed.
Input DataSets	Domo data sources that will be available to use in the Jupyter workspace
Output DataSets	Data sets that will be available to write data as part of the Jupyter processing

 

Editing a Workspace

1. From the Jupyter Workspaces list, navigate to the workspace you wish to edit.

2. Hover to the right of the workspace. The Manage Workspace menu displays. 

3. Select the Manage Workspace menu and choose Edit.

The Edit Jupyter Workspace modal displays.

4. Add new specifications to the workspace and select Save. 

The workspace is updated with new details. 

 

Deleting a Workspace

1. From the Jupyter Workspaces list, navigate to the workspace you wish to delete. 

2. Hover to the right of the workspace. The Manage Workspace menu displays. 

3. Select Manage Workspace menu and choose Delete.

The Delete Test? modal displays.

4. Confirm that you wish to delete the workspace by selecting the Delete button. 

The workspace is deleted. This action cannot be undone. 

 

Jupyter Basics 

Reading Data

Data can be read into a Jupyter Notebook using the domojupyter library. This library provides useful functionality to interact with Domo within Jupyter. Before data can be read into a Jupyter Notebook using the domojupyter library, your workspace must have an input data source. 

If you are creating a new workspace, see the earlier section, Getting Started.

If your workspace already exists but doesn't have an input data source, follow the steps below: 

1. In Jupyter Workspaces, locate your workspace from the list. You can search and filter by owner. 

2. Hover over your workspace. The Manage Workspace  menu displays. 

3. Select the Manage Workspace menu and choose Edit. 

The Edit Jupyter Workspace modal displays. 

4. In the Edit Jupyter Workspace view, select Add Input DataSet and choose a data source to be read into a Jupyter Notebook. 

5. After choosing a DataSet, select Save. 

The workspace has an input DataSet, and data can now be read into a Jupyter Notebook using the domojupyter library. All DataSets used in the workspace are located in Data > DataSets. 

6. Use the domo.read_dataframe command to read data from a Domo data source into your Jupyter Notebook for further processing and analysis.

See the example below: 

 

Writing Data

Data can be written back to a Domo data source using the domojupyter library. This library provides useful functionality to interact with Domo within Jupyter. Before data can be written into a Jupyter Notebook using the domojupyter library, your workspace must have an output data source. 

If you are creating a new workspace, see the earlier section, Getting Started.

If your workspace already exists but doesn't have an output data source, follow the steps below: 

1. In Jupyter Workspaces, locate your workspace from the list. You can search and filter by owner. 

2. Hover over your workspace. The Manage Workspace  menu displays. 

3. Select Manage Workspacemenu and choose Edit. 

The Edit Jupyter Workspace modal displays. 

4. In the Edit Jupyter Workspace view, select Add Output DataSet and choose a data source to be written into a Jupyter Notebook. 

5. After choosing a DataSet, select Save. 

The workspace has an output DataSet, and data can now be written into a Jupyter Notebook using the domojupyter library. All DataSets used in the workspace are located in Data > DataSets. 

6. Use the domo.write_dataframecommand to write results back to the Domo Data sources configured for the Jupyter workspace.

See the example below: 

 

Installing and Using Libraries

Libraries can be installed in the Jupyter workspace by opening a terminal and executing the appropriate commands. An example command to install the Seaborn library is conda install seaborn -y. Once installed, these libraries can be imported and used within the Jupyter Notebooks.

See the example below:

 

Scheduling a Notebook

Schedule a Jupyter notebook by selecting the dataflow option within the Juptyer Notebook interface.

 

Once selected, you will be presented with the option to create a Jupyter Dataflow. Enter a Name, Description, and desired scheduling options. Select Save.

Once saved, you will see the data flow in the DataFlows section of the Data Center.

You can navigate back to the Jupyter Workspace by Selecting the Edit in Jupyter button in the details view of the data flow.

 

Usage Monitoring

You can monitor Jupyter usage within Domo.

1. Navigate to Admin > Features

3. Select Jupyter > Usage.

Here, you can see the total hours used across workspaces.

 

To see the detailed usage, select a workspace. You will see associated data flows and the server specs being used. 

 

Experiencing issues? See Jupyter Troubleshooting Guide or contact support@domo.com for assistance. 

 

FAQs

Can I install Plugins?

Plugins are not currently supported.

Can I share Jupyter Notebooks?

It is on our roadmap to provide this functionality. You can transfer ownership of a workspace, which provides a limited sharing capability.

Can I trial the feature?

Please contact your Customer Success Manager or Account Executive for the most up-to-date information on the trial experience of Jupyter.

Does Domo offer data science services?      

Yes, we have a team of experienced data scientists ready to help as needed. Please contact your Account Executive for more information.

How do I change the kernel? 

First, you must stop the workspace. You may then change the kernel in the Domo workspace edit view.

Can I reset the kernel?

Yes. When you hover over a workspace, the Manage Workspaces menu displays. Choose Restore Kernel Defaults from the options. 

I'm experiencing issues integrating third-party libraries.

Domo's Jupyter integration allows users to install third-party libraries. However, Domo will not natively support every third-party library. Some third-party libraries may require more effort on the user's end to install, configure, and troubleshoot.