EngRAD: Multivariate Weather Measurements Across England

<h1>EngRAD Dataset</h1> <p>The EngRAD dataset contains measurements of 5 different weather variables collected at 487 grid points in England from 2018 to 2020.</p> <h2>Dataset Overview</h2> <p>Data has been provided by <a href="https://open-meteo.com/" target="_blank" rel="noopener">Open-Meteo</a>&nbsp;and licensed under <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank" rel="noopener">Attribution 4.0 International (CC BY 4.0)</a>. The numerical weather prediction model used to generate the data is ECMWF IFS, which has a spatial resolution of 9 km. The grid points are located in correspondence with cities. Each point is associated with location information such as geographic coordinates, elevation, closest city, and the county to which it belongs. The physical variables collected are:</p> <ol> <li>Air temperature at 2 meters above ground (&deg;C)</li> <li>Relative humidity at 2 meters above ground (%)</li> <li>Summation of total precipitation (rain, showers, snow) during the preceding hour (mm)</li> <li>Total cloud cover (%)</li> <li>Global horizontal irradiance (W/m&sup2;)</li> </ol> <p>These variables are typically of interest in applications related to solar radiation, such as solar power production.</p> <h3>Channels:</h3> <p>The <code>data.h5</code> file contains two tables, accessible by the following keys:</p> <ul> <li> <p><strong><code>data</code></strong>: Contains the measurements for each point across different weather variables.</p> <ul> <li><strong><code>temperature_2m</code></strong>: Air temperature at 2 meters above ground (&deg;C). Instantaneous measurement.</li> <li><strong><code>relative_humidity_2m</code></strong>: Relative humidity at 2 meters above ground (%). Instantaneous measurement.</li> <li><strong><code>precipitation</code></strong>: Total precipitation (rain, showers, snow) sum of the preceding hour (mm). Preceding hour sum.</li> <li><strong><code>cloud_cover</code></strong>: Total cloud cover as an area fraction (%). Instantaneous measurement.</li> <li><strong><code>shortwave_radiation</code></strong>: Global horizontal irradiation (GHI) (W/m&sup2;). Preceding hour mean.</li> </ul> </li> </ul> <p>&nbsp;</p> <ul> <li> <p><strong><code>metadata</code></strong>: Contains the following detailed information for each point:</p> <ul> <li><code><strong>city</strong></code>: The name of the city where the measurement point is located.</li> <li><code><strong>county</strong></code>: The county in which the city is situated.</li> <li><code><strong>admin_name</strong></code>: The administrative name associated with the city or region.</li> <li><code><strong>lat</strong></code>: The latitude coordinate of the measurement point.</li> <li><code><strong>lon</strong></code>: The longitude coordinate of the measurement point.</li> <li><code><strong>elevation</strong></code>: The elevation (in meters) above sea level at the measurement point.</li> <li><code><strong>population</strong></code>: The population of the city where the measurement point is located.</li> </ul> </li> </ul> <h3>Dataset size</h3> <ul> <li><strong>Time steps</strong>: 26304</li> <li><strong>Points</strong>: 487</li> <li><strong>Channels</strong>: 5</li> <li><strong>Sampling rate</strong>: 1 hour</li> <li><strong>Missing values</strong>: 0.00%</li> </ul> <h3>Credits</h3> <p>This dataset has been introduced in the paper:</p> <p>Ivan Marisca, Cesare Alippi, and Filippo Maria Bianchi. "Graph-based forecasting with missing data through spatiotemporal downsampling."&nbsp;<em>Proceedings of the 41st International Conference on Machine Learning</em>, PMLR 235:34846-34865, 2024.</p> <p>Please consider citing the paper if you use the dataset for your research.</p> <pre><code>@inproceedings{marisca2024graph,</code><br><code>&nbsp; title &nbsp; &nbsp; = {Graph-based Forecasting with Missing Data through Spatiotemporal Downsampling},</code><br><code>&nbsp; author &nbsp; &nbsp;= {Marisca, Ivan and Alippi, Cesare and Bianchi, Filippo Maria},</code><br><code>&nbsp; booktitle = {Proceedings of the 41st International Conference on Machine Learning},</code><br><code>&nbsp; pages &nbsp; &nbsp; = {34846--34865},</code><br><code>&nbsp; year &nbsp; &nbsp; &nbsp;= {2024},</code><br><code>&nbsp; volume &nbsp; &nbsp;= {235},</code><br><code>&nbsp; series &nbsp; &nbsp;= {Proceedings of Machine Learning Research},</code><br><code>&nbsp; publisher = {PMLR}</code><br><code>}</code></pre>