Forecasting

Forecasting predicts future values based on historical data patterns using exponential smoothing models. It automatically detects seasonality, trends, and cycles in time-series data. Forecasts include confidence intervals showing prediction uncertainty. Tableau handles missing data and multiple measures, making sophisticated forecasting accessible without statistical expertise.

To show the forecast, perform the following steps.

1. Connect Tableau with the Kyvos semantic model.

2. Using the Kyvos semantic model, create a time-series view by dragging the month(orderdate) and sales amount into the view, as shown below.

   Open

   

3. Go to the analytics tab and drag the forecast into the visualization pane.

   Open

   

4. Tableau extends the line with forecasted values (shaded region).

   Open

   

5. Right-click forecast area > Forecast Options.

   Open

   

6. Adjust forecast length (months/quarters to predict)

7. Set confidence intervals (80%, 95%, etc.)

8. View model description for quality metrics and click on Ok.

9. The user will receive the forecasting according to the configuration.

   Open