autogluon/fev_datasets

Forecast evaluation datasets

This repository contains time series datasets that can be used for evaluation of univariate & multivariate forecasting models.

The main focus of this repository is on datasets that reflect real-world forecasting scenarios, such as those involving covariates, missing values, and other practical complexities.

The datasets follow a format that is compatible with the fev package.

Data format and usage

Each dataset satisfies the following schema:

• each dataset entry (=row) represents a single univariate or multivariate time series
• each entry contains
  • 1/ a field of type Sequence(timestamp) that contains the timestamps of observations
  • 2/ at least one field of type Sequence(float) that can be used as the target time series or dynamic covariates
  • 3/ a field of type string that contains the unique ID of each time series
• all fields of type Sequence have the same length

Datasets can be loaded using the 🤗 datasets library.

import datasets

ds = datasets.load_dataset("autogluon/fev_datasets", "epf_electricity_de", split="train")
ds.set_format("numpy")  # sequences returned as numpy arrays

Example entry in the epf_electricity_de dataset

>>> ds[0]
{'id': 'DE',
 'timestamp': array(['2012-01-09T00:00:00.000000', '2012-01-09T01:00:00.000000',
        '2012-01-09T02:00:00.000000', ..., '2017-12-31T21:00:00.000000',
        '2017-12-31T22:00:00.000000', '2017-12-31T23:00:00.000000'],
       dtype='datetime64[us]'),
 'target': array([34.97, 33.43, 32.74, ...,  5.3 ,  1.86, -0.92], dtype=float32),
 'Ampirion Load Forecast': array([16382. , 15410.5, 15595. , ..., 15715. , 15876. , 15130. ],
       dtype=float32),
 'PV+Wind Forecast': array([ 3569.5276,  3315.275 ,  3107.3076, ..., 29653.008 , 29520.33  ,
        29466.408 ], dtype=float32)}

For more details about the dataset format and usage, check out the fev documentation on GitHub.

Dataset statistics

Disclaimer: These datasets have been converted into a unified format from external sources. Please refer to the original sources for licensing and citation terms. We do not claim any rights to the original data.

config	freq	# items	# obs	# dynamic cols	# static cols	source	citation
epf_electricity_be	h	1	157248	3	0	https://zenodo.org/records/4624805	[1]
epf_electricity_de	h	1	157248	3	0	https://zenodo.org/records/4624805	[1]
epf_electricity_fr	h	1	157248	3	0	https://zenodo.org/records/4624805	[1]
epf_electricity_np	h	1	157248	3	0	https://zenodo.org/records/4624805	[1]
epf_electricity_pjm	h	1	157248	3	0	https://zenodo.org/records/4624805	[1]
favorita_store_sales	D	1782	12032064	4	6	https://www.kaggle.com/competitions/store-sales-time-series-forecasting	[2]
favorita_transactions	D	54	273456	3	5	https://www.kaggle.com/competitions/store-sales-time-series-forecasting	[2]
m5_with_covariates	D	30490	428849460	9	5	https://www.kaggle.com/competitions/m5-forecasting-accuracy	[3]
proenfo_bull	h	41	2877216	4	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_cockatoo	h	1	105264	6	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_covid19	h	1	223384	7	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_gfc12_load	h	11	867108	2	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_gfc14_load	h	1	35040	2	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_gfc17_load	h	8	280704	2	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_hog	h	24	2526336	6	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_pdb	h	1	35040	2	0	https://github.com/Leo-VK/EnFoAV	[4]
proenfo_spain	h	1	736344	21	0	https://github.com/Leo-VK/EnFoAV	[4]

Publications using these datasets

• "ChronosX: Adapting Pretrained Time Series Models with Exogenous Variables"