Creating a univariate forecast

Edit online

Create a BAU forecast from actual activity levels using historical data from one variable to predict future values. This forecast is a BAU outlook that is generated before any abatement programs are applied.

Before you begin

Open the Baseline Forecast worksheet in the Envizi Emissions Forecasting & Planning workspace.

Confirm that Actuals are loaded and locked to preserve data integrity. Daily refresh mirrors IBM® ESG Suite updates.

Ensure that the exploration grid displays data from the Forecast cube.

About this task

Planning Analytics Workspace automatically selects a time-series model, such as exponential smoothing or ARIMA, based on data characteristics. The model generates forecast values and confidence intervals, which you can view in the forecast preview.

Tip:
  • Save forecast results to the Forecast version. Using other versions causes an error.
  • Use the Seasonality spreading method to distribute values realistically across months. With this method, Actuals are used to determine the spread of values taking into consideration underlying, long-term trends from recurring or seasonal fluctuations. Forecasting at the leaf or branch level usually provides better accuracy than forecasting at highly aggregated roots.

Procedure

  1. Set filters in the exploration grid:
    Set Version to Actuals and Scenario to Baseline.
  2. Use dimensions to specify what data you want to forecast:
    For example, set Measure to CO₂ or set Utility Type to Electricity.
  3. Select a location to forecast, such as Grand Central.
  4. From the top-level menu, choose Forecast > Univariate Forecast.
  5. In the side preview panel, choose the start and end periods for your forecast.
  6. In the Advanced section on the preview panel, select the data spreading method by setting Spread Forecast Values to Relative proportional (Seasonality). This spreads forecast values based on the seasonality detected in historical data.
  7. Next, select the version where you want to save the predicted values. Set Dimension to Version, Hierarchy to Version, Member to Forecast to save your predicted values to the Forecast version.
  8. Click Preview to visualize results and review model accuracy statistics.
  9. Click Forecast to save values to the Forecast version of the Forecast cube.
  10. Confirm that new forecast values appear in the grid under Scenario = Baseline.

Results

Historic and forecast data is displayed together in the Actuals + Forecast version to represent the BAU scenario.

What to do next

You can run the same forecast across other rows and save the values to the Forecast cube. To run forecasts in batch mode, press Shift to select multiple consecutive rows. Press Control to select non-consecutive rows.