Datasets¶
The Datasets tab lists all datasets currently linked to the selected Use Case by you and other team members. To access this tab, open a Use Case and click Datasets.
From this tab, you can:
Element | Description | |
---|---|---|
1 | Add new | Add a dataset, experiment, or notebook to your Use Case, or create a new Use Case. |
2 | Search | Search for a specific dataset. |
3 | Sort | Sort the dataset columns. |
4 | More options | Click More options to interact with a dataset:
|
View exploratory data insights¶
While a dataset is being registered in Workbench, DataRobot also performs exploratory data analysis (EDA1)—analyzing and profiling every feature to detect feature types, automatically transform date-type features, and assess feature quality. Once registration is complete, you can explore the information uncovered while computing EDA1.
Public preview
Support for dynamic datasets in Workbench is on by default.
When this feature is enabled:
- Datasets added via a data connection will be registered as dynamic datasets in the Data Registry and Use Case.
- Dynamic datasets added via a connection will be available for selection in the Data Registry.
- DataRobot will pull a new live sample when viewing Exploratory Data Insights for dynamic datasets.
Feature flag: Enable Dynamic Datasets in Workbench
To view exploratory data insights:
- In a Use Case, navigate to the Datasets tab.
-
Click the More options icon next to the dataset you want to view and select Explore. Alternatively, click the dataset name to view its insights.
-
For each feature in the dataset, DataRobot displays various feature details, including a histogram and summary statistics.
-
To drill down into a specific feature, click its histogram chart along the top.
View wrangling recipe SQL¶
If you performed wrangling on a registered dataset, you can see the final compiled form of the operations executed by the cloud data platform after pushdown. To view the SQL queries executed by the cloud data platform, select the dataset and click Recipe SQL.
Feature lists¶
Public preview
Support for feature lists in Workbench is on by default.
Feature flag: Enable Feature Lists in Workbench Preview
After adding a dataset to your Use Case, DataRobot generates feature lists as part of EDA. Feature lists control the subset of features that DataRobot uses to build models and make predictions. Each model has a feature list associated to it.
You might want to use feature lists to:
- Remove features that cannot be used in the model for any reason, for example, a feature that is causing target leakage.
- Make predictions faster by removing unimportant features (i.e., ones that don't improve the model's performance).
You can use one of the automatically created lists—Informative and Raw—or create a custom feature list.
View feature lists¶
Before setting up an experiment, use exploratory data insights to explore different feature lists before choosing the appropriate one to use for modeling.
To explore insights for a feature list:
-
In the Datasets tab, click the More options icon next to the dataset you want to view and select Explore. Alternatively, click the dataset name.
-
To access your feature lists, click the dropdown at the top of the page and select an available feature list. The preview updates to show only the features in the selected list.
Create a feature list¶
To create a custom feature list:
-
While exploring a dataset, click the dropdown at the top of the page and select + New feature list. This opens the Features view.
-
Select the box next to each feature you want to include in your custom list. Then, click Create feature list.
-
Enter a name and description (optional) for the new feature list.
-
Click Save changes. You can now access the new feature list in the dropdown.
Next steps¶
From here, you can:
- Add more data.
- Perform data wrangling for datasets added via a data connection.
- Use the dataset to set up an experiment and start modeling.