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`](https://github.com/autogluon/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`](https://huggingface.co/docs/datasets/en/index) library.

```python
import datasets

ds = datasets.load_dataset("autogluon/fev_datasets", "epf_de", split="train")
ds.set_format("numpy") # sequences returned as numpy arrays
```
Example entry in the `epf_de` dataset
```python
>>> 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](https://github.com/autogluon/fev?tab=readme-ov-file#tutorials).

## 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. Unless otherwise specified, the datasets are provided only for research purposes.

| config | freq | # items | median length | # obs | # dynamic cols | # static cols | source | citation |
|:---------------------------|:-------|:----------|:----------------|:------------|-----------------:|----------------:|:---------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------|
| `ETT_15T` | 15min | 2 | 69,680 | 975,520 | 7 | 0 | https://github.com/zhouhaoyi/ETDataset | [[1]](https://arxiv.org/abs/2012.07436) |
| `ETT_1D` | D | 2 | 724 | 10,136 | 7 | 0 | https://github.com/zhouhaoyi/ETDataset | [[1]](https://arxiv.org/abs/2012.07436) |
| `ETT_1H` | h | 2 | 17,420 | 243,880 | 7 | 0 | https://github.com/zhouhaoyi/ETDataset | [[1]](https://arxiv.org/abs/2012.07436) |
| `ETT_1W` | W-SUN | 2 | 103 | 1,442 | 7 | 0 | https://github.com/zhouhaoyi/ETDataset | [[1]](https://arxiv.org/abs/2012.07436) |
| `LOOP_SEATTLE_1D` | D | 323 | 365 | 117,895 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `LOOP_SEATTLE_1H` | h | 323 | 8,760 | 2,829,480 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `LOOP_SEATTLE_5T` | 5min | 323 | 105,120 | 33,953,760 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `M_DENSE_1D` | D | 30 | 730 | 21,900 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `M_DENSE_1H` | h | 30 | 17,520 | 525,600 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `SZ_TAXI_15T` | 15min | 156 | 2,976 | 464,256 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `SZ_TAXI_1H` | h | 156 | 744 | 116,064 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[2]](https://arxiv.org/abs/2304.14343) |
| `australian_tourism` | QE-DEC | 89 | 36 | 3,204 | 1 | 0 | https://robjhyndman.com/publications/hierarchical-tourism/ | [[3]](https://doi.org/10.1016/j.ijforecast.2008.07.004) |
| `bizitobs_l2c_1H` | h | 1 | 2,664 | 18,648 | 7 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `bizitobs_l2c_5T` | 5min | 1 | 31,968 | 223,776 | 7 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `boomlet_1062` | 5min | 1 | 16,384 | 344,064 | 21 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1209` | 5min | 1 | 16,384 | 868,352 | 53 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1225` | min | 1 | 16,384 | 802,816 | 49 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1230` | 5min | 1 | 16,384 | 376,832 | 23 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1282` | min | 1 | 16,384 | 573,440 | 35 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1487` | 5min | 1 | 16,384 | 884,736 | 54 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1631` | 30min | 1 | 10,463 | 418,520 | 40 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1676` | 30min | 1 | 10,463 | 1,046,300 | 100 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1855` | h | 1 | 5,231 | 272,012 | 52 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_1975` | h | 1 | 5,231 | 392,325 | 75 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_2187` | h | 1 | 5,231 | 523,100 | 100 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_285` | min | 1 | 16,384 | 1,228,800 | 75 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_619` | min | 1 | 16,384 | 851,968 | 52 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_772` | min | 1 | 16,384 | 1,097,728 | 67 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `boomlet_963` | min | 1 | 16,384 | 458,752 | 28 | 6 | https://huggingface.co/datasets/Datadog/BOOM | [[5]](https://arxiv.org/abs/2505.14766) |
| `ecdc_ili` | W-SUN | 25 | 201 | 4,797 | 1 | 0 | https://github.com/EU-ECDC/Respiratory_viruses_weekly_data/blob/main/data/snapshots/2025-08-08_ILIARIRates.csv | |
| `entsoe_15T` | 15min | 6 | 175,292 | 6,310,512 | 6 | 0 | https://data.open-power-system-data.org/time_series/2020-10-06 | [[6]](https://doi.org/10.25832/time_series/2020-10-06) |
| `entsoe_1H` | h | 6 | 43,822 | 1,577,592 | 6 | 0 | https://data.open-power-system-data.org/time_series/2020-10-06 | [[6]](https://doi.org/10.25832/time_series/2020-10-06) |
| `entsoe_30T` | 30min | 6 | 87,645 | 3,155,220 | 6 | 0 | https://data.open-power-system-data.org/time_series/2020-10-06 | [[6]](https://doi.org/10.25832/time_series/2020-10-06) |
| `epf_be` | h | 1 | 52,416 | 157,248 | 3 | 0 | https://zenodo.org/records/4624805 | [[7]](https://doi.org/10.1016/j.apenergy.2021.116983) |
| `epf_de` | h | 1 | 52,416 | 157,248 | 3 | 0 | https://zenodo.org/records/4624805 | [[7]](https://doi.org/10.1016/j.apenergy.2021.116983) |
| `epf_fr` | h | 1 | 52,416 | 157,248 | 3 | 0 | https://zenodo.org/records/4624805 | [[7]](https://doi.org/10.1016/j.apenergy.2021.116983) |
| `epf_np` | h | 1 | 52,416 | 157,248 | 3 | 0 | https://zenodo.org/records/4624805 | [[7]](https://doi.org/10.1016/j.apenergy.2021.116983) |
| `epf_pjm` | h | 1 | 52,416 | 157,248 | 3 | 0 | https://zenodo.org/records/4624805 | [[7]](https://doi.org/10.1016/j.apenergy.2021.116983) |
| `ercot_1D` | D | 8 | 6,452 | 51,616 | 1 | 0 | https://github.com/ourownstory/neuralprophet-data/tree/main/datasets_raw/energy | |
| `ercot_1H` | h | 8 | 154,872 | 1,238,976 | 1 | 0 | https://github.com/ourownstory/neuralprophet-data/tree/main/datasets_raw/energy | |
| `ercot_1M` | ME | 8 | 211 | 1,688 | 1 | 0 | https://github.com/ourownstory/neuralprophet-data/tree/main/datasets_raw/energy | |
| `ercot_1W` | W-SUN | 8 | 921 | 7,368 | 1 | 0 | https://github.com/ourownstory/neuralprophet-data/tree/main/datasets_raw/energy | |
| `favorita_stores_1D` | D | 1,579 | 1,688 | 10,661,408 | 4 | 6 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `favorita_stores_1M` | ME | 1,579 | 54 | 255,798 | 3 | 6 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `favorita_stores_1W` | W-SUN | 1,579 | 240 | 1,136,880 | 3 | 6 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `favorita_transactions_1D` | D | 51 | 1,688 | 258,264 | 3 | 5 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `favorita_transactions_1M` | ME | 51 | 54 | 5,508 | 2 | 5 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `favorita_transactions_1W` | W-SUN | 51 | 240 | 24,480 | 2 | 5 | https://www.kaggle.com/competitions/store-sales-time-series-forecasting | [[8]](https://www.kaggle.com/competitions/store-sales-time-series-forecasting/overview/citation) |
| `fred_md_2025` | MS | 1 | 798 | 100,548 | 126 | 0 | https://www.stlouisfed.org/research/economists/mccracken/fred-databases | [[9]](https://doi.org/10.20955/wp.2015.012) |
| `fred_qd_2025` | QS-DEC | 1 | 266 | 65,170 | 245 | 0 | https://www.stlouisfed.org/research/economists/mccracken/fred-databases | [[10]](https://doi.org/10.20955/wp.2020.005) |
| `gvar` | QS-OCT | 33 | 178 | 52,866 | 9 | 0 | https://data.mendeley.com/datasets/kfp5fhgkvf/1 | [[11]](https://doi.org/10.17863/CAM.104755) |
| `hermes` | W-MON | 10,000 | 261 | 5,220,000 | 2 | 2 | https://github.com/etidav/HERMES | [[12]](https://arxiv.org/abs/2202.03224) |
| `hierarchical_sales_1D` | D | 118 | 1,825 | 215,350 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `hierarchical_sales_1W` | W-WED | 118 | 260 | 30,680 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `hospital_admissions_1D` | D | 8 | 1,731 | 13,846 | 1 | 0 | https://www.kaggle.com/datasets/datasetengineer/riyadh-hospital-admissions-dataset-20202024 | [[13]](https://doi.org/10.34740/kaggle/dsv/9992619) |
| `hospital_admissions_1W` | W-SUN | 8 | 246 | 1,968 | 1 | 0 | https://www.kaggle.com/datasets/datasetengineer/riyadh-hospital-admissions-dataset-20202024 | [[13]](https://doi.org/10.34740/kaggle/dsv/9992619) |
| `hospital` | ME | 767 | 84 | 64,428 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `jena_weather_10T` | 10min | 1 | 52,704 | 1,106,784 | 21 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `jena_weather_1D` | D | 1 | 366 | 7,686 | 21 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `jena_weather_1H` | h | 1 | 8,784 | 184,464 | 21 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `kdd_cup_2022_10T` | 10min | 134 | 35,279 | 47,273,860 | 10 | 0 | https://aistudio.baidu.com/competition/detail/152/0/task-definition | [[14]](https://arxiv.org/abs/2208.04360) |
| `kdd_cup_2022_1D` | D | 134 | 243 | 325,620 | 10 | 0 | https://aistudio.baidu.com/competition/detail/152/0/task-definition | [[14]](https://arxiv.org/abs/2208.04360) |
| `kdd_cup_2022_30T` | 30min | 134 | 11,758 | 15,755,720 | 10 | 0 | https://aistudio.baidu.com/competition/detail/152/0/task-definition | [[14]](https://arxiv.org/abs/2208.04360) |
| `m5_1D` | D | 30,490 | 1,810 | 428,849,460 | 9 | 5 | https://www.kaggle.com/competitions/m5-forecasting-accuracy | [[15]](https://doi.org/10.1016/j.ijforecast.2021.11.013) |
| `m5_1M` | ME | 30,490 | 58 | 13,805,685 | 9 | 5 | https://www.kaggle.com/competitions/m5-forecasting-accuracy | [[15]](https://doi.org/10.1016/j.ijforecast.2021.11.013) |
| `m5_1W` | W-SUN | 30,490 | 257 | 60,857,703 | 9 | 5 | https://www.kaggle.com/competitions/m5-forecasting-accuracy | [[15]](https://doi.org/10.1016/j.ijforecast.2021.11.013) |
| `proenfo_bull` | h | 41 | 17,544 | 2,877,216 | 4 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_cockatoo` | h | 1 | 17,544 | 105,264 | 6 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_gfc12` | h | 11 | 39,414 | 867,108 | 2 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_gfc14` | h | 1 | 17,520 | 35,040 | 2 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_gfc17` | h | 8 | 17,544 | 280,704 | 2 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_hog` | h | 24 | 17,544 | 2,526,336 | 6 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `proenfo_pdb` | h | 1 | 17,520 | 35,040 | 2 | 0 | https://github.com/Leo-VK/EnFoAV | [[16]](https://doi.org/10.48550/arXiv.2307.07191) |
| `redset_15T` | 15min | 126 | 8,640 | 1,052,371 | 1 | 1 | https://github.com/amazon-science/redset/ | [[17]](https://www.amazon.science/publications/why-tpc-is-not-enough-an-analysis-of-the-amazon-redshift-fleet) |
| `redset_1H` | h | 138 | 2,160 | 283,070 | 1 | 1 | https://github.com/amazon-science/redset/ | [[17]](https://www.amazon.science/publications/why-tpc-is-not-enough-an-analysis-of-the-amazon-redshift-fleet) |
| `redset_5T` | 5min | 118 | 25,920 | 2,960,408 | 1 | 1 | https://github.com/amazon-science/redset/ | [[17]](https://www.amazon.science/publications/why-tpc-is-not-enough-an-analysis-of-the-amazon-redshift-fleet) |
| `restaurant` | D | 817 | 296 | 294,568 | 1 | 4 | https://www.kaggle.com/c/recruit-restaurant-visitor-forecasting | [[18]](www.kaggle.com/competitions/recruit-restaurant-visitor-forecasting/overview/citation) |
| `rohlik_orders_1D` | D | 7 | 1,197 | 115,650 | 15 | 0 | https://www.kaggle.com/competitions/rohlik-orders-forecasting-challenge | [[19]](www.kaggle.com/competitions/rohlik-orders-forecasting-challenge/overview/citation) |
| `rohlik_orders_1W` | W-SUN | 7 | 170 | 15,316 | 14 | 0 | https://www.kaggle.com/competitions/rohlik-orders-forecasting-challenge | [[19]](www.kaggle.com/competitions/rohlik-orders-forecasting-challenge/overview/citation) |
| `rohlik_sales_1D` | D | 5,390 | 1,046 | 74,413,935 | 15 | 7 | https://www.kaggle.com/competitions/rohlik-sales-forecasting-challenge-v2 | [[20]](https://www.kaggle.com/competitions/rohlik-sales-forecasting-challenge-v2/overview/citation) |
| `rohlik_sales_1W` | W-SUN | 5,243 | 150 | 10,516,770 | 15 | 7 | https://www.kaggle.com/competitions/rohlik-sales-forecasting-challenge-v2 | [[20]](https://www.kaggle.com/competitions/rohlik-sales-forecasting-challenge-v2/overview/citation) |
| `rossmann_1D` | D | 1,115 | 942 | 7,352,310 | 7 | 10 | https://www.kaggle.com/competitions/rossmann-store-sales | [[21]](www.kaggle.com/competitions/rossmann-store-sales/overview/citation) |
| `rossmann_1W` | W-SUN | 1,115 | 133 | 889,770 | 6 | 10 | https://www.kaggle.com/competitions/rossmann-store-sales | [[21]](www.kaggle.com/competitions/rossmann-store-sales/overview/citation) |
| `solar_1D` | D | 137 | 365 | 50,005 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `solar_1W` | W-FRI | 137 | 52 | 7,124 | 1 | 0 | https://huggingface.co/datasets/Salesforce/GiftEval | [[4]](https://arxiv.org/abs/2410.10393) |
| `solar_with_weather_15T` | 15min | 1 | 198,600 | 1,986,000 | 10 | 0 | https://www.kaggle.com/datasets/samanemami/renewable-energy-and-weather-conditions | |
| `solar_with_weather_1H` | h | 1 | 49,648 | 496,480 | 10 | 0 | https://www.kaggle.com/datasets/samanemami/renewable-energy-and-weather-conditions | |
| `uci_air_quality_1D` | D | 1 | 389 | 5,057 | 13 | 0 | https://archive.ics.uci.edu/dataset/360/air+quality | [[22]](https://doi.org/10.24432/C59K5F) |
| `uci_air_quality_1H` | h | 1 | 9,357 | 121,641 | 13 | 0 | https://archive.ics.uci.edu/dataset/360/air+quality | [[22]](https://doi.org/10.24432/C59K5F) |
| `uk_covid_nation_1D` | D | 4 | 729 | 41,216 | 14 | 0 | https://www.kaggle.com/datasets/happyadam73/uk-covid19-dashboard-data-sqlite-compressed | |
| `uk_covid_nation_1W` | W-SUN | 4 | 105 | 5,936 | 14 | 0 | https://www.kaggle.com/datasets/happyadam73/uk-covid19-dashboard-data-sqlite-compressed | |
| `uk_covid_utla_1D` | D | 214 | 721 | 308,786 | 2 | 0 | https://www.kaggle.com/datasets/happyadam73/uk-covid19-dashboard-data-sqlite-compressed | |
| `uk_covid_utla_1W` | W-SUN | 214 | 104 | 44,448 | 2 | 0 | https://www.kaggle.com/datasets/happyadam73/uk-covid19-dashboard-data-sqlite-compressed | |
| `us_consumption_1M` | MS | 31 | 792 | 24,552 | 1 | 0 | https://apps.bea.gov/iTable/?reqid=19&step=3&isuri=1&nipa_table_list=2017&categories=underlying | [[23]](https://doi.org/10.1016/j.ijforecast.2016.04.005) |
| `us_consumption_1Q` | QE-DEC | 31 | 262 | 8,122 | 1 | 0 | https://apps.bea.gov/iTable/?reqid=19&step=3&isuri=1&nipa_table_list=2017&categories=underlying | [[23]](https://doi.org/10.1016/j.ijforecast.2016.04.005) |
| `us_consumption_1Y` | YE-DEC | 31 | 64 | 1,984 | 1 | 0 | https://apps.bea.gov/iTable/?reqid=19&step=3&isuri=1&nipa_table_list=2017&categories=underlying | [[23]](https://doi.org/10.1016/j.ijforecast.2016.04.005) |
| `walmart` | W-FRI | 2,936 | 143 | 4,609,143 | 11 | 4 | https://www.kaggle.com/competitions/walmart-recruiting-store-sales-forecasting | [[24]](www.kaggle.com/competitions/walmart-recruiting-store-sales-forecasting/overview/citation) |
| `world_co2_emissions` | YE-DEC | 191 | 60 | 11,460 | 1 | 0 | https://www.kaggle.com/datasets/ulrikthygepedersen/co2-emissions-by-country | |
| `world_life_expectancy` | YE-DEC | 237 | 74 | 17,538 | 1 | 0 | https://www.kaggle.com/datasets/nafayunnoor/global-life-expectancy-data-1950-2023 | [[25]](https://ourworldindata.org/life-expectancy#article-citation) |
| `world_tourism` | YE-DEC | 178 | 21 | 3,738 | 1 | 0 | https://www.kaggle.com/datasets/bushraqurban/tourism-and-economic-impact | [[26]](https://www.worldbank.org/en/archive/using-the-archives/terms-of-use-reproduction-and-citation) |

## Citation
If you find these datasets useful in your work, please cite the following [paper](https://arxiv.org/abs/2509.26468)
```
@article{shchur2025fev,
title={{fev-bench}: A Realistic Benchmark for Time Series Forecasting},
author={Shchur, Oleksandr and Ansari, Abdul Fatir and Turkmen, Caner and Stella, Lorenzo and Erickson, Nick and Guerron, Pablo and Bohlke-Schneider, Michael and Wang, Yuyang},
year={2025},
eprint={2509.26468},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```

## Publications using these datasets

- ["ChronosX: Adapting Pretrained Time Series Models with Exogenous Variables"](https://arxiv.org/abs/2503.12107)