> ## Documentation Index
> Fetch the complete documentation index at: https://nixtlaverse.nixtla.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Feature Engineering | UtilsForecast

> Create exogenous regressors for your models

### `fourier`

```python theme={null}
fourier(df, freq, season_length, k, h=0, id_col='unique_id', time_col='ds')
```

Compute fourier seasonal terms for training and forecasting

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `df` | <code>pandas or polars DataFrame</code> | Dataframe with ids, times and values for the exogenous regressors. | *required* |
| `freq` | <code>[str](#str) or [int](#int)</code> | Frequency of the data. Must be a valid pandas or polars offset alias, or an integer. | *required* |
| `season_length` | <code>[int](#int)</code> | Number of observations per unit of time. Ex: 24 Hourly data. | *required* |
| `k` | <code>[int](#int)</code> | Maximum order of the fourier terms | *required* |
| `h` | <code>[int](#int)</code> | Forecast horizon. Defaults to 0. | <code>0</code> |
| `id_col` | <code>[str](#str)</code> | Column that identifies each serie. Defaults to 'unique\_id'. | <code>'unique\_id'</code> |
| `time_col` | <code>[str](#str)</code> | Column that identifies each timestep, its values can be timestamps or integers. Defaults to 'ds'. | <code>'ds'</code> |

**Returns:**

| Type | Description |
| - | - |
| <code>[Tuple](#typing.Tuple)\[[DFType](#utilsforecast.compat.DFType), [DFType](#utilsforecast.compat.DFType)]</code> | tuple\[pandas or polars DataFrame, pandas or polars DataFrame]: A tuple containing the original DataFrame with the computed features and DataFrame with future values. |

### `trend`

```python theme={null}
trend(df, freq, h=0, id_col='unique_id', time_col='ds')
```

Add a trend column with consecutive integers for training and forecasting

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `df` | <code>pandas or polars DataFrame</code> | Dataframe with ids, times and values for the exogenous regressors. | *required* |
| `freq` | <code>[str](#str) or [int](#int)</code> | Frequency of the data. Must be a valid pandas or polars offset alias, or an integer. | *required* |
| `h` | <code>[int](#int)</code> | Forecast horizon. Defaults to 0. | <code>0</code> |
| `id_col` | <code>[str](#str)</code> | Column that identifies each serie. Defaults to 'unique\_id'. | <code>'unique\_id'</code> |
| `time_col` | <code>[str](#str)</code> | Column that identifies each timestep, its values can be timestamps or integers. Defaults to 'ds'. | <code>'ds'</code> |

**Returns:**

| Type | Description |
| - | - |
| <code>[Tuple](#typing.Tuple)\[[DFType](#utilsforecast.compat.DFType), [DFType](#utilsforecast.compat.DFType)]</code> | tuple\[pandas or polars DataFrame, pandas or polars DataFrame]: A tuple containing the original DataFrame with the computed features and DataFrame with future values. |

### `time_features`

```python theme={null}
time_features(df, freq, features, h=0, id_col='unique_id', time_col='ds')
```

Compute timestamp-based features for training and forecasting

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `df` | <code>pandas or polars DataFrame</code> | Dataframe with ids, times and values for the exogenous regressors. | *required* |
| `freq` | <code>[str](#str) or [int](#int)</code> | Frequency of the data. Must be a valid pandas or polars offset alias, or an integer. | *required* |
| `features` | <code>list of str or callable</code> | Features to compute. Can be string aliases of timestamp attributes or functions to apply to the times. | *required* |
| `h` | <code>[int](#int)</code> | Forecast horizon. Defaults to 0. | <code>0</code> |
| `id_col` | <code>[str](#str)</code> | Column that identifies each serie. Defaults to 'unique\_id'. | <code>'unique\_id'</code> |
| `time_col` | <code>[str](#str)</code> | Column that identifies each timestep, its values can be timestamps or integers. Defaults to 'ds'. | <code>'ds'</code> |

**Returns:**

| Type | Description |
| - | - |
| <code>[Tuple](#typing.Tuple)\[[DFType](#utilsforecast.compat.DFType), [DFType](#utilsforecast.compat.DFType)]</code> | tuple\[pandas or polars DataFrame, pandas or polars DataFrame]: A tuple containing the original DataFrame with the computed features and DataFrame with future values. |

### `future_exog_to_historic`

```python theme={null}
future_exog_to_historic(df, freq, features, h=0, id_col='unique_id', time_col='ds')
```

Turn future exogenous features into historic by shifting them `h` steps.

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `df` | <code>pandas or polars DataFrame</code> | Dataframe with ids, times and values for the exogenous regressors. | *required* |
| `freq` | <code>[str](#str) or [int](#int)</code> | Frequency of the data. Must be a valid pandas or polars offset alias, or an integer. | *required* |
| `features` | <code>list of str</code> | Features to be converted into historic. | *required* |
| `h` | <code>[int](#int)</code> | Forecast horizon. Defaults to 0. | <code>0</code> |
| `id_col` | <code>[str](#str)</code> | Column that identifies each serie. Defaults to 'unique\_id'. | <code>'unique\_id'</code> |
| `time_col` | <code>[str](#str)</code> | Column that identifies each timestep, its values can be timestamps or integers. Defaults to 'ds'. | <code>'ds'</code> |

**Returns:**

| Type | Description |
| - | - |
| <code>[Tuple](#typing.Tuple)\[[DFType](#utilsforecast.compat.DFType), [DFType](#utilsforecast.compat.DFType)]</code> | tuple\[pandas or polars DataFrame, pandas or polars DataFrame]: A tuple containing the original DataFrame with the computed features and DataFrame with future values. |

### `pipeline`

```python theme={null}
pipeline(df, features, freq, h=0, id_col='unique_id', time_col='ds')
```

Compute several features for training and forecasting

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `df` | <code>pandas or polars DataFrame</code> | Dataframe with ids, times and values for the exogenous regressors. | *required* |
| `features` | <code>list of callable</code> | List of features to compute. Must take only df, freq, h, id\_col and time\_col (other arguments must be fixed). | *required* |
| `freq` | <code>[str](#str) or [int](#int)</code> | Frequency of the data. Must be a valid pandas or polars offset alias, or an integer. | *required* |
| `h` | <code>[int](#int)</code> | Forecast horizon. Defaults to 0. | <code>0</code> |
| `id_col` | <code>[str](#str)</code> | Column that identifies each serie. Defaults to 'unique\_id'. | <code>'unique\_id'</code> |
| `time_col` | <code>[str](#str)</code> | Column that identifies each timestep, its values can be timestamps or integers. Defaults to 'ds'. | <code>'ds'</code> |

**Returns:**

| Type | Description |
| - | - |
| <code>[Tuple](#typing.Tuple)\[[DFType](#utilsforecast.compat.DFType), [DFType](#utilsforecast.compat.DFType)]</code> | tuple\[pandas or polars DataFrame, pandas or polars DataFrame]: A tuple containing the original DataFrame with the computed features and DataFrame with future values. |


## Related topics

- [Feature engineering | MLForecast](/mlforecast/feature_engineering.html.md)
- [Feature engineering | StatsForecast](/statsforecast/src/feature_engineering.html.md)
- [Generating features](/statsforecast/docs/how-to-guides/generating_features.html.md)
- [utilsforecast](/utilsforecast/index.html.md)
- [Quick start (distributed)](/mlforecast/docs/getting-started/quick_start_distributed.html.md)