Dark Mode
European Multi-City Weather Features Dataset
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset provides daily weather observations for 18 European cities and locations, collected between 2000 and 2010. It is designed for machine learning and deep learning training and teaching purposes, offering a robust foundation for various analytical tasks. The dataset is suitable for classification, regression, and forecasting, presenting enough complexity to illustrate realistic challenges such as overfitting and unbalanced data, while remaining intuitively accessible.
Columns
The primary data file,
weather_prediction_dataset.csv, contains a range of weather features. An optional file, weather_prediction_picnic_labels.csv, includes boolean labels indicating whether daily conditions were suitable for a picnic.Below are the main columns with their descriptions and physical units after conversion for machine learning suitability:
- _temp_mean: Mean daily temperature in 1 °C
- _temp_max: Maximum daily temperature in 1 °C
- _temp_min: Minimum daily temperature in 1 °C
- _cloud_cover: Cloud cover in oktas
- _wind_gust: Wind gust in 1 m/s
- _wind_speed: Wind speed in 1 m/s
- _humidity: Humidity in 1 %
- _pressure: Sea level pressure in 1000 hPa
- _global_radiation: Global radiation in 100 W/m2
- _precipitation: Daily precipitation in 10 mm
- _sunshine: Sunshine hours in 0.1 hours
Distribution
The dataset is provided in a tabular, comma-separated CSV format. The main data file,
weather_prediction_dataset.csv, comprises 165 variables (features) recorded over 3654 daily observations. A metadata.txt file is also available, offering further details on data processing and variable descriptions. The total data is approximately 4.71 MB.Usage
This dataset is ideally suited for academic research, educational initiatives, and practical machine learning applications. It can be utilised for:
- Developing and testing classification models for weather events.
- Building regression models to predict specific weather parameters.
- Training forecasting models for future weather conditions.
- Demonstrating and addressing machine learning challenges like overfitting and handling unbalanced data.
Coverage
The dataset's geographic scope covers 18 distinct European cities and locations: Basel (Switzerland), Budapest (Hungary), Dresden, Düsseldorf, Kassel, München (all Germany), De Bilt and Maastricht (the Netherlands), Heathrow (UK), Ljubljana (Slovenia), Malmo and Stockholm (Sweden), Montélimar, Perpignan and Tours (France), Oslo (Norway), Roma (Italy), and Sonnblick (Austria). The time range spans from 2000 to 2010, encompassing 3654 daily observations. Data for certain weather variables, such as cloud cover, wind speed, humidity, and global radiation, were included wherever available for the selected locations.
License
Attribution 4.0 International (CC BY 4.0)
Who Can Use It
This dataset is valuable for a wide range of users, including:
- Data Scientists: For building and refining predictive models.
- Machine Learning Engineers: For training and evaluating deep learning architectures.
- Academics and Researchers: For studying climate patterns and atmospheric science.
- Educators: For teaching machine learning concepts and practical data analysis.
- Students: For projects involving classification, regression, and time-series analysis.
Dataset Name Suggestions
- European Daily Weather Predictions (2000-2010)
- EU Climate Assessment Weather Data
- European Multi-City Weather Features Dataset
- ML-Ready European Weather Observations
Attributes
Original Data Source: European Multi-City Weather Features Dataset
QUALITY