MeteoGalEus Dataset

This one of the official repositories of the MeteoGalEus Dataset, presented at 15th Language Resources and Evaluation Conference (LREC) in Palma, Mallorca (Spain, 11-16 May 2026). If you use the MeteoGalEus dataset in any project, please quote the paper where it is described (see the Citation section at the bottom for full citation in .bib format):

A. Vivel-Couso, N. Z. Pramata, D. Robredo, A. Soroa, J.M. Alonso-Moral, "MeteoGalEus: An Iberian Multilingual Weather Dataset in Galician, Euskera, and Spanish." Proceedings of the 15th Language Resources and Evaluation Conference (LREC2026).

This is a simplify version that only contains the dataset. For more information, please check the original repo in GitLab: ainhoa.vivel/meteogaleus.

Index

• Repository
• Description
  • Dataset variables
• Examples
  • Galicia (GL+ES)
  • Basque Country (EU+ES)
• Data Analysis
• License
• Citation

Repository

The repository only contains the MeteoGalEus dataset.

Back to Index

Description

MeteoGalEus dataset is a multilingual, cross-regional weather corpus that combines structured meteorological data and human-written forecasts. It covers three languages: Spanish, Galician, and Euskera. The entire dataset is distributed in JSON format, facilitating easy integration and analysis. Each entry corresponds to a single day and a specific region, comprising both meteorological predictions and forecast texts authored by professional meteorologists. The data spans 33 to 36 cities per region, providing comprehensive coverage. This dataset supports a range of research applications, including cross-regional weather analysis, multilingual natural language generation, climate trend studies, evaluation of large language models, and investigations into low-resource languages, particularly Euskera and Galician.

NOTE: More information about the dataset can be found in the paper.

Back to Index

Dataset variables

Each entry in the dataset corresponds to a single day and contains both textual forecasts and meteorological measurements for a specific regional agency:

• id. A unique identifier that preserves trace to the original sources (e.g., euskalmet_2020-11-25 or 3158). This way the dataset is endowed with traceability, allowing users to easily locate the raw data in the source agency. This ensures reproducibility and enables researchers to cross-check or enrich the dataset with additional information from the original records.
• from. The source agency (MeteoGalicia or Euskalmet).
• autonomous_community. The geographic area (Galicia or the Basque Country).
• date. The date of the forecast in ISO 8601 format (e.g., 2020-11-25).
• forecast. The textual pieces of information written by human experts (i.e., professional meteorologists affiliated to each agency). Forecasts are available in Spanish (stored under the key ES) as well as in the co-official language of the corresponding autonomous community: Euskera (EU) for the Basque Country and Galician (GL) for Galicia.
• meteorological_data. A structured list of measurements for multiple locations within the corresponding autonomous community. This field is represented as an array that typically contains data for 33 to 36 cities, depending on availability for a given day. For each city:
  • zone. The subregion name (e.g., Lugo).
  • location. The city name (e.g., Bilbao).
  • temperature. Daily maximum (max) and minimum (min) temperatures.
  • wind. A list of measurements for three time periods (morning, afternoon, night), including direction (cardinal points) and speed (e.g., calm).
  • state_of_the_sky. The state of the sky for a location, including cloud coverage and precipitation type for different times of day (morning, afternoon, night)

Back to Index

Examples

In this section we provide illustrative examples from our multilingual MeteoGalEus dataset, highlighting the structure and content of the data. Each example demonstrates how meteorological predictions are paired with professional forecasts in different languages. We present examples separately for the Galician and Euskera examples to showcase the multilingual aspect of the corpus.

Back to Index

Galicia (GL+ES)

This subsection presents examples of entries in Galician, covering both structured meteorological data (temperature, wind, state of the sky) and corresponding human-written forecasts in Galician and Spanish. The examples illustrate how professional meteorologists describe daily weather conditions for multiple cities within the Galician region, demonstrating the alignment between the data and textual descriptions.

NOTE: In this example, for clarity and brevity, the meteorological_data field displays only a single city.

{
    "id": 3159,
    "from": "MeteoGalicia",
    "autonomous_community": "Galicia",
    "date": "2020-11-26",
    "forecast": {
      "ES": "El jueves a borrasca se retirará hacia el sur, dejando a Galicia en una zona intermedia entre altas y bajas presiones, con vientos de componente este. Así, se esperan cielos con intervalos de nubes y claros, con posibilidad de chubascos, más probables cuanto más al sur. Las temperaturas mínimas continuaron sin cambios, mientras que las máximas ascenderán moderadamente. Los vientos soplarán de componente este, flojos en el interior y moderados en la costa.",
      "GL": "Hoxe xoves a borrasca retirarase cara ó sur, deixando a Galicia nunha zona intermedia entre altas e baixas presións, con ventos de compoñente leste. Así, agárdanse ceos con intervalos de nubes e claros, con posibilidade de chuvascos, máis probables canto máis ó sur. As temperaturas mínimas continuaron sen cambios ou lixeiros descensos, mentres que as máximas ascenderán moderadamente. Os ventos soprarán de compoñente leste, frouxos no interior e moderados na costa."
    },
    "meteorological_data": [
      {
        "zone": "Ourense",
        "location": "A Limia",
        "temperature": {
          "max": 9.666666666666666,
          "min": 6.333333333333333
        },
        "wind": [
          {
            "period": "mañana",
            "direction": "este",
            "speed": "viento débil"
          },
          {
            "period": "noche",
            "direction": "este",
            "speed": "viento débil"
          },
          {
            "period": "tarde",
            "direction": "este",
            "speed": "viento débil"
          }
        ],
        "state_of_the_sky": {
          "morning": "chubascos débiles",
          "night": "nublado",
          "afternoon": "chubascos débiles"
        }
      },
    ]
}

Back to Index

Basque Country (EU+ES)

This subsection presents examples of entries in Basque Country, covering both structured meteorological data (temperature, wind, state of the sky) and corresponding human-written forecasts in Basque and Spanish. The examples illustrate how professional meteorologists describe daily weather conditions for multiple cities within the Basque Country, demonstrating the alignment between the data and textual descriptions.

Note: In this example, for clarity and brevity, the meteorological_data field displays only a single city.

{
  "id": "euskalmet_2020-11-25",
  "from": "Euskalmet",
  "autonomous_community": "Basque Country",
  "date": "2020-11-25",
  "forecast": {
    "ES": "Durante lo que resta de día tendremos el cielo poco nuboso en todas partes salvo en el Valle del Ebro y comarcas cercanas, donde se espesará de nuevo la niebla. El viento del sureste dejará algunas rachas muy fuertes en zonas expuestas y hará que las temperaturas nocturnas vuelvan a subir, sobre todo en el interior.",
    "EU": "Gainerako orduetan zerua hodei gutxirekin egongo da denean, Ebro bailaran izan ezik; han behe-hodeiak eta lainoa izango dira nagusi. Hego-ekialdeko haizea bizi ibiliko da, bolada oso gogorrekin haizeguneetan. Tenperatura minimoek gora egingo dute berriro, batez ere barnealdean, kostaldeko baxuenek berdintsu jarraituko dute ziurrenik."
  },
  "meteorological_data": [
    {
      "zone": "Bizkaia",
      "location": "Bilbao",
      "temperature": {
        "max": 11.6,
        "min": 11.6
      },
      "wind": [
        {
          "period": "mañana",
          "direction": "sur",
          "speed": "calma"
        },
        {
          "period": "tarde",
          "direction": "sur",
          "speed": "calma"
        },
        {
          "period": "noche",
          "direction": "sur",
          "speed": "calma"
        }
      ],
      "state_of_the_sky": {
        "morning": "despejado",
        "afternoon": "despejado",
        "night": "despejado"
      }
    },
  ]
}

Back to Index

License

The MeteoGalEus dataset is licensed under the terms and conditions of the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).

Back to Index

Citation

@inproceedings{vivel-couso-etal-2026-meteogaleus,
    title = {{M}eteo{G}al{E}us: {A}n {I}berian {M}ultilingual {W}eather {D}ataset in {G}alician, {E}uskera, and {S}panish},
    author = {Vivel-Couso, A.  and Pramata, N. Z. and Robredo, D. and Soroa, A. and Alonso-Moral, J.M.},
    year = 2026,
    month = may,
    booktitle = {Proceedings of the 15th Language Resources and Evaluation Conference},
    publisher = {ELRA Language Resources Association},
    address = {Palau de Congressos de Palma, Mallorca (Spain) and virtual meeting},
    pages = {},
    url = {},
    abstract = {This paper introduces MeteoGalEus, a multilingual weather dataset that combines meteorological observations from two Spanish regional agencies, Euskalmet and MeteoGalicia. The dataset contains daily records spanning 4 years and 3 months, with aligned observations for both sources. MeteoGalEus captures key meteorological variables including temperature, wind and state of the sky. The dataset is provided in a structured format, facilitating data analysis and integration, with textual forecasts available in the official languages for each region (i.e., Galician and Spanish for MeteoGalicia; Euskera and Spanish for Euskalmet). By merging and harmonizing data from two regional agencies, MeteoGalEus is a unique resource for cross-regional weather analysis and multilingual climate studies. This dataset is suited for tasks requiring high-quality, aligned, and standardized weather data across multiple languages and regions. We conducted baseline experiments using LLaMA-based models in both zero-shot and fine-tuned settings to illustrate the use of MeteoGalEus for natural language generation (NLG). Fine-tuning led to consistent improvements across all metrics, with BERTScore increasing from 0.68 to 0.79, ROUGE from 0.20 to 0.35, and BLEU from 0.02 to 0.17 in the best-performing model. The experiments show how MeteoGalEus can be taken as a benchmark for multilingual and cross-regional NLG tasks.}
}

Back to Index