Dark Mode
ML-Ready European Daily Weather Features
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This weather data tracks daily meteorological features designed specifically for use in machine learning and deep learning applications. It is an ideal resource for educational purposes, helping users explore concepts like classification, regression, and time-series forecasting, while also tackling real-world challenges such as overfitting and unbalanced data. The collection includes detailed daily metrics for 18 European cities, sourced from the European Climate Assessment & Dataset (ECA&D).
Columns
The data fields capture essential weather metrics for 18 distinct cities over the defined period. The core features included, prefixed by the city name (e.g., BASEL_), are:
- DATE: The specific date of the observation.
- MONTH: The numerical month of the observation.
- _temp_mean, _temp_max, _temp_min: Mean, maximum, and minimum daily temperatures, measured in degrees Celsius (°C).
- _cloud_cover: The extent of cloud coverage, measured in oktas.
- _global_radiation: Incoming solar radiation, scaled in units of 100 W/m².
- _humidity: Relative humidity, expressed as a percentage (%).
- _pressure: Atmospheric pressure, scaled in units of 1000 hPa.
- _precipitation: Daily rainfall or accumulated precipitation, scaled in units of 10 mm.
- _sunshine: Duration of direct sunshine, scaled in units of 0.1 hours.
- _wind_speed, _wind_gust: Measurements for wind speed and maximum wind gust, both in metres per second (m/s).
Distribution
The primary data is distributed in CSV format, specifically the
weather_prediction_dataset.csv file, which is approximately 2.8 MB in size. The structure includes daily records, with over 3,600 individual observations covering the full decade for each location. Supplementary files are also available, including a slimmed-down version, a location map, and optional labels for "picnic suitability" (True/False classification).Usage
This dataset is perfect for training predictive models, enabling users to:
- Develop and benchmark weather forecasting models.
- Perform classification tasks, such as predicting the likelihood of "picnic weather".
- Execute regression analyses on specific meteorological variables.
- Educate students or practitioners on machine learning principles using complex, real-world climate data.
Coverage
The data spans 18 cities across Europe. The temporal coverage is a full decade, running from 2000 through to 2010. The data captures daily measurements for this entire period, providing a continuous time-series suitable for longitudinal analysis.
License
Creative Commons Attribution 4.0 International (CC BY 4.0)
Who Can Use It
- Data Scientists: For building advanced forecasting models and experimenting with various time-series algorithms.
- Academics and Educators: To teach core machine learning concepts like feature engineering, classification, and dealing with imbalanced data.
- Students: For capstone projects or assignments focused on climate data analysis and prediction.
- Developers: For creating applications that require historical European weather trends.
Dataset Name Suggestions
- Decadal European Climate Data
- ML-Ready European Daily Weather Features
- Weather Prediction Training Dataset (2000-2010)
- Eighteen Cities Daily Weather Record
Attributes
Original Data Source: ML-Ready European Daily Weather Features
QUALITYLoading...