> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.kodelabs.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Forecasting
> Configure and train energy usage forecasts for short-term consumption projections in EnerG
Train a forecast configuration to project near-term energy usage for a building from interval meter data. In the [building view](/products/energ/get-started/navigation#energy-modeling-submenu), expand **Energy Modeling** in the left sidebar and open **Forecast**. Results also appear on the [Building Energy dashboard](/products/energ/analytics/building-dashboards-energy#forecast) when you select **Forecast** under **Energy Modeling**. For how forecasting extends trained models, see [Energy modeling and M\&V](/products/energ/concepts/energy-modeling#forecasting).
Forecasts rely on interval reads with enough history for the model to learn daily and weekly patterns. Confirm meter completeness on the [Meters](/products/energ/buildings/meters) page before you train.
## Open Forecast
From [Portfolio Home](/products/energ/analytics/benchmarking), open a building in the building view.
In the left sidebar, expand **Energy Modeling**, then click **Forecast**.
## Manage configurations
The **Forecast** page lists every forecast configuration for the building. The page subtitle describes the default horizon: forecast energy usage trends over the coming week. Click **+ Create Configuration** to start the three-step wizard. Use the **Actions** menu on a row to edit or delete an existing configuration.
| Column | What it shows |
| --------------- | -------------------------------------------------------------------------------------------------- |
| **Name** | Configuration label. The **Default** tag marks the configuration EnerG uses on forecast dashboards |
| **Aggregation** | Rollup level, such as **Hourly** |
| **Meters** | Count of meters included |
| **Status** | Training state, such as **Succeeded** |
| **Actions** | Row menu for edit, delete, or related tasks |
## Create a forecast configuration
Click **+ Create Configuration** to open the wizard. Progress appears in the left sidebar across three steps: **Basic Setup**, **Data Selection**, and **Summary**.
### Step 1: Basic setup
Define the configuration name and whether it is the default forecast for the building.
| Field | Purpose |
| --------------- | ---------------------------------------------------------------- |
| **Name** | Label shown in lists and dashboard filters |
| **Default** | When checked, EnerG selects this configuration on forecast views |
| **Description** | Optional notes for your team |
Click **Next** to continue.
### Step 2: Data selection
Choose aggregation, training start date, interval meters, and accuracy thresholds.
| Field | Purpose |
| ------------------------ | ---------------------------------------------------------------------------------------------------------------------------- |
| **Aggregation** | Granularity for the model, such as **Hourly** |
| **Start Date** | Optional date to limit training data. Leave empty to use all available history |
| **Meter list** | Search and filter by **Meter Scope** and **Service Type**, then select meters |
| **Point selected** | For interval meters, pick the point type EnerG uses for the forecast, such as an electric accumulator |
| **Threshold Percentage** | **Model Metrics** uses Mean Absolute Percentage Error (MAPE) as the threshold, or **Set the Threshold** for a custom percent |
Many teams use **Hourly** aggregation when they need a week-ahead operational view. Match the point type to the same signal you use in [meter comparisons](/products/energ/data-quality/meter-comparisons).
### Step 3: Summary
Review trained model metrics before you finish.
The summary confirms how many meters trained (**Meters Included** versus **Meters Selected**) and the **Last Trained** date. It also reports **CVRMSE**, **MAPE**, and **R Squared** fit. For metric definitions and thresholds, see [Model accuracy thresholds](/products/energ/reference/calculations-and-parameters#model-accuracy).
Click **Finish** to save the configuration. Projected consumption appears on forecast charts after processing completes.
## Read forecast results
Open **Forecast** from the Building Energy dashboard secondary menu to compare projected consumption to actual use for recent and upcoming periods. Charts update after the scheduled sync cycle, similar to [M\&V analysis](/products/energ/energy-modeling/mv-analysis) results.
## When to use forecasting
| Use case | Approach |
| -------------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
| Near-term operations | Train an **Hourly** configuration to see usage trends over the coming week |
| Peak load planning | Compare projected peaks to actual interval data after major equipment changes |
| Budget checks | Align forecast horizons with monthly budget periods on [Building budget overview](/products/energ/finance/budget-overview) |
| Savings validation | Pair forecasts with [M\&V](/products/energ/energy-modeling/mv-models) when you need both adjusted baselines and forward projections |
## When training fails
Forecast runs can fail when interval history is sparse or correlation is weak. Verify the meter has continuous reads, widen the training window, or change the selected point before you retry.
If MAPE or CVRMSE exceed your threshold, try a different point type or confirm the meter is not duplicated in [metrics](/products/energ/buildings/meters#include-or-exclude-from-metrics).
## Next steps
Review forecast charts without leaving the Energy dashboard.
Configure baselines and savings models alongside forecasting.
Interpret savings and accuracy metrics after each model run completes.
See how M\&V, forecasting, and IPMVP practices fit together in EnerG.