Skip to content

Click in-app to access the full platform documentation for your version of DataRobot.

Workflow overview: Build time series models

To begin, upload a dataset that meets the time series dataset requirements. If DataRobot's data quality assessment detects an irregular time step in the dataset, you can use the time series data prep tool to correct issues.

If DataRobot detects one or more time variables in your dataset, the time-aware modeling functionality becomes active. Following is the general workflow, but see the more detailed page on date/time modeling for complete information.

Note

Even if your data contains time features, time series forecasting mode may be disabled if the data contains irregular time units or non-unique time stamps. See the section describing the time series data prep tool for potential solutions.

After setting a target feature:

  1. Click the Set up time-aware modeling link to get started.

  2. From the dropdown, select the primary date/time feature:

  3. Select the time-aware approach you would like to apply:

    • Use automated machine learning (OTV) when your data is time-relevant but you are not forecasting (instead, you are predicting the target value on each individual row). Use this if you have single event data, such as patient intake or loan defaults.
    • Use automated time series forecasting when you want to forecast multiple future values of the target (for example, predicting sales for each day next week). Use this to extrapolate future values in a continuous sequence.
    • Use automated time series nowcasting when you want to use modeling to determine current values.
  4. If you selected time series and DataRobot detected series data, set the series ID for multiseries modeling.

  5. Customize forecast windows (Feature Derivation Window (FDW) and Forecast Window (FW)) to configure how DataRobot derives features for the modeling dataset. Also consider:

    • Employing the data prep tool.
    • Setting features as known in advance (to be used unlagged when making predictions).
    • Including a calendar file.
  6. To modify the settings used for modeling (date/time format, training window, validation length, etc.), scroll down and expand Show advanced options. See the full documentation for more information.

    To modify the multiseries series identifier, open the time series advanced options.

  7. Once set, choose a modeling mode and press Start.

Once the modeling process begins, DataRobot analyzes the target and implements time series best practices. DataRobot also creates time-based features to use in the different blueprints. Consider working with the following:

  • Drill down on variables in the Data tab to view distributions and trends.
  • Work with the time series feature lists used for modeling.
  • Use the Prediction Preview to explore the predictions, download the values, and visualize the "confidence" factor with the prediction interval.

Investigate models

Once your models are built, the following tabs are available from the Leaderboard to help with model evaluation:

Tab Availability
Accuracy Over Time OTV; additional options for time series and multiseries
Forecast vs Actual Time series, multiseries
Series Insights Multiseries
Stability OTV, time series, multiseries
Forecasting Accuracy Time series, multiseries
Anomaly Over Time Anomaly detection: OTV, time series, multiseries
Anomaly Assessment Anomaly detection: time series, multiseries

Updated December 6, 2021
Back to top