To model with exogenous features, you have two options: 1. Use historical exogenous variables: include these variables in the DataFrame you pass to the forecast method 2. Use future exogenous variables: include these variables in the DataFrame you pass to the forecast method and provide the future values of these exogenous features over the forecast horizon using the X_df parameter.

import pandas as pd
from nixtla import NixtlaClient
nixtla_client = NixtlaClient(
    # defaults to os.environ.get("NIXTLA_API_KEY")
    api_key = 'my_api_key_provided_by_nixtla'
)

👍 Use an Azure AI endpoint

To use an Azure AI endpoint, set the base_url argument:

nixtla_client = NixtlaClient(base_url="you azure ai endpoint", api_key="your api_key")

​
1. Historical exogenous variables

# Read data
df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short-with-ex-vars.csv')

# Forecast
forecast_df = nixtla_client.forecast(
    df=df, 
    h=24,
    id_col='unique_id',
    target_col='y',
    time_col='ds',
    # Add the columns of `df` that will be considered as historical
    hist_exog_list=['Exogenous1', 'Exogenous2', 'day_0', 'day_1', 'day_2', 'day_3', 'day_4', 'day_5', 'day_6']
)

​
2. Future exogenous variables

# Read data
import numpy as np
df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short-with-ex-vars.csv')

# Load the future value of exogenous variables over the forecast horizon
future_ex_vars_df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short-future-ex-vars.csv')

# Forecast
forecast_df = nixtla_client.forecast(
    df=df, 
    X_df=future_ex_vars_df, 
    h=24,
    id_col='unique_id',
    target_col='y',
    time_col='ds'
)

​
3. Historical and future exogenous variables

# Read data
df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short-with-ex-vars.csv')

# Load the future value of exogenous variables over the forecast horizon
future_ex_vars_df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short-future-ex-vars.csv')

# We will only use 2 exogenous of future_ex_vars_df
future_ex_vars_df = future_ex_vars_df[["unique_id", "ds", "Exogenous1", "Exogenous2"]]
# To pass historical exogenous variables, we need to add the list of columns
# in the `hist_exog_list` as follows.

# Forecast
forecast_df = nixtla_client.forecast(
    df=df, 
    X_df=future_ex_vars_df, 
    h=24,
    id_col='unique_id',
    target_col='y',
    time_col='ds',
    # Add the columns of `df` that will be considered as historical
    hist_exog_list=['day_0', 'day_1', 'day_2', 'day_3', 'day_4', 'day_5', 'day_6']
)

📘 Available models in Azure AI

If you use an Azure AI endpoint, set model="azureai"

nixtla_client.detect_anomalies(..., model="azureai")

For the public API, two models are supported: timegpt-1 and timegpt-1-long-horizon.

By default, timegpt-1 is used. See this tutorial for details on using timegpt-1-long-horizon.

For more details on using exogenous features with TimeGPT, read our in-depth tutorials on Exogenous variables and on Categorical variables.