- This data-set is intended for use in research on Machine Learning and Time-Series Prediction.
- The weather-data covers the period between 1980-2018 for five Danish cities: Aalborg, Aarhus, Esbjerg, Odense and Roskilde.
- Each data-point measures the temperature (Celcius), the barometric pressure (hecto-pascal or milli-bar), the wind-speed (meters per second), and the wind-direction (angular degrees).
- Some data is partially missing, e.g. periods of the barometric pressure in some cities.
- Some errors have been discovered in the data, e.g. a few temperature-measurements are clearly wrong.
The raw weather-data was originally obtained from the National Climatic Data Center (NCDC) in USA.
The raw data-file had some formatting problems that had to be corrected manually. This archive contains a cleaned-up version of the data-file, that is easier to load either as a CSV-file or as a so-called pickle-file for use in Pandas with Python.
The archived tar-ball is automatically downloaded and extracted by using the weather.py module for Python.
According to the instructions on NCDC's web-site, this data can only be used and redistributed for non-commercial purposes.
