# Generate forecasts from statistical data

Statistical forecasting applies historical trends and seasonality to generate forecasts without manual input. The prediction feature enables forward‑looking analysis by deriving future values from historical patterns. It supports the configuration of confidence interval, seasonality, and growth factor, and allows top-down or bottom-up approaches to apply forecasts across hierarchies. Results can be explored visually or in tabular form.

### Prerequisites

* A forecast measure with blank initial values.
* The planning sheet must contain at least 2 years of historical data 

### Configure statistical forecasts

{% hint style="info" %}
The forecast configuration can be saved to a profile. Select the profile to reuse a predefined configuration.
{% endhint %}

1. Select the cell for which to create a forecast.
2. Go to **Model** > **Predict**.
3. Select More options (…) and then select **Create New Profile** and enter the profile name. If a profile already exists, select the profile.

<figure><img src="/files/lC5gY6PtaDewR6bmnmcg" alt=""><figcaption></figcaption></figure>

4. Select the historical date range based on which predictions will be generated.
5. Select the forecast date range.&#x20;

{% hint style="info" %}
The forecast range can extend beyond the open period. Extending the range is useful when the open forecast period is short (for example, Q1 only), but predicted values are needed over a longer period to assess accuracy and get a better perspective.
{% endhint %}

6. Your data may show repeating patterns over time, known as seasonality. Select the seasonality type:

* If your data repeats at one regular interval, choose **Single Seasonality**.\
  For example, sales that repeat every month.
* If your data repeats at more than one time level, choose **Multi‑Seasonality**.\
  This applies when the data shows patterns at multiple intervals, such as a quarterly trend (Q1, Q2, Q3, Q4), and a monthly pattern within each quarter.

7. Set the **Confidence (%)** to indicate the level of certainty that the actual value will fall within the predicted range.
8. Set the **Growth Factor (%)** that defines the expected rate at which the predicted values increase or decrease over time.
9. Assign the **Evaluation** method:&#x20;
   * To enter the forecast for a parent cell and distribute it to child categories, select **Top Down**.
   * To enter forecasts at the child level and aggregate them to the parent category, select B**ottom Up**.
10. Select **Run Forecast**.
11. Preview the predicted values in a graph:
    * Historical data points are plotted in grey.
    * Predicted data points are plotted in green.
    * The green reference band indicates the confidence interval.

<figure><img src="/files/pLWFEX0TvVSXPDa1ltIK" alt="" width="375"><figcaption></figcaption></figure>

12. Select **Table** to review the actual predicted values in tabular format.

<figure><img src="/files/wF0epEFgE655SQkkSr5N" alt="" width="375"><figcaption></figcaption></figure>

13. Select **Save Forecast** after parameters are adjusted to create an accurate prediction.

<figure><img src="/files/wwMcLsLWrsHCAKeOYQS1" alt=""><figcaption></figcaption></figure>

14. Predictions are generated for the selected measure and row category. As the Bottom-Up approach is used, values are automatically aggregated to the parent levels.

<figure><img src="/files/rOExOeed9mhpbdh9xT4s" alt=""><figcaption></figcaption></figure>

### Locking selections

When the **Predict** panel opens, the selected cell can still be changed by selecting a different visible cell on the canvas. The updated row and column selection is reflected in the Predict panel.

When the selection is locked, selecting other cells does not change the row or column used for prediction.

<figure><img src="/files/FTTInYKxkajfzVqE4OMR" alt=""><figcaption></figcaption></figure>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.fabricplan.com/planning-sheets/how-tos/forecast-data-to-predict-future-trends/generate-forecasts-from-statistical-data.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.