> ## 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.

> Forecasting utilities

# utilsforecast

## Install

### PyPI

```sh theme={null}
pip install utilsforecast
```

### Conda

```sh theme={null}
conda install -c conda-forge utilsforecast
```

***

## How to use

### Generate synthetic data

```python theme={null}
from utilsforecast.data import generate_series

series = generate_series(3, with_trend=True, static_as_categorical=False)
series
```

```
|     | unique_id | ds         | y          |
|-----|-----------|------------|------------|
| 0   | 0         | 2000-01-01 | 0.422133   |
| 1   | 0         | 2000-01-02 | 1.501407   |
| 2   | 0         | 2000-01-03 | 2.568495   |
| 3   | 0         | 2000-01-04 | 3.529085   |
| 4   | 0         | 2000-01-05 | 4.481929   |
| ... | ...       | ...        | ...        |
| 481 | 2         | 2000-06-11 | 163.914625 |
| 482 | 2         | 2000-06-12 | 166.018479 |
| 483 | 2         | 2000-06-13 | 160.839176 |
| 484 | 2         | 2000-06-14 | 162.679603 |
| 485 | 2         | 2000-06-15 | 165.089288 |
```

***

### Plotting

```python theme={null}
from utilsforecast.plotting import plot_series

fig = plot_series(series, plot_random=False, max_insample_length=50, engine='matplotlib')
fig.savefig('imgs/index.png', bbox_inches='tight')
```

<img src="https://mintlify.s3.us-west-1.amazonaws.com/nixtla/utilsforecast/docs/mintlify/imgs/index.png" alt="" />

<img src="https://mintcdn.com/nixtla/w6Yby9GibAjo68ns/utilsforecast/imgs/index.png?fit=max&auto=format&n=w6Yby9GibAjo68ns&q=85&s=6a55f132b507a653955bfdac4136c570" alt="" width="1697" height="711" data-path="utilsforecast/imgs/index.png" />

***

### Preprocessing

```python theme={null}
from utilsforecast.preprocessing import fill_gaps

serie = series[series['unique_id'].eq(0)].tail(10)
# drop some points
with_gaps = serie.sample(frac=0.5, random_state=0).sort_values('ds')
with_gaps
```

Example output with missing dates:

```
|     | unique_id | ds         | y         |
|-----|-----------|------------|-----------|
| 213 | 0         | 2000-08-01 | 18.543147 |
| 214 | 0         | 2000-08-02 | 19.941764 |
| 216 | 0         | 2000-08-04 | 21.968733 |
| 220 | 0         | 2000-08-08 | 19.091509 |
| 221 | 0         | 2000-08-09 | 20.220739 |
```

```python theme={null}
fill_gaps(with_gaps, freq='D')
```

Returns:

```
|     | unique_id | ds         | y         |
|-----|-----------|------------|-----------|
| 0   | 0         | 2000-08-01 | 18.543147 |
| 1   | 0         | 2000-08-02 | 19.941764 |
| 2   | 0         | 2000-08-03 | NaN       |
| 3   | 0         | 2000-08-04 | 21.968733 |
| 4   | 0         | 2000-08-05 | NaN       |
| 5   | 0         | 2000-08-06 | NaN       |
| 6   | 0         | 2000-08-07 | NaN       |
| 7   | 0         | 2000-08-08 | 19.091509 |
| 8   | 0         | 2000-08-09 | 20.220739 |
```

***

### Evaluating

```python theme={null}
from functools import partial
import numpy as np

from utilsforecast.evaluation import evaluate
from utilsforecast.losses import mape, mase
```

```python theme={null}
valid = series.groupby('unique_id').tail(7).copy()
train = series.drop(valid.index)

rng = np.random.RandomState(0)
valid['seas_naive'] = train.groupby('unique_id')['y'].tail(7).values
valid['rand_model'] = valid['y'] * rng.rand(valid['y'].shape[0])

daily_mase = partial(mase, seasonality=7)

evaluate(valid, metrics=[mape, daily_mase], train_df=train)
```

```
|     | unique_id | metric | seas_naive | rand_model |
|-----|-----------|--------|------------|------------|
| 0   | 0         | mape   | 0.024139   | 0.440173   |
| 1   | 1         | mape   | 0.054259   | 0.278123   |
| 2   | 2         | mape   | 0.042642   | 0.480316   |
| 3   | 0         | mase   | 0.907149   | 16.418014  |
| 4   | 1         | mase   | 0.991635   | 6.404254   |
| 5   | 2         | mase   | 1.013596   | 11.365040  |
```

***