UK RSE Conference 2022 Walkthrough - MLOps for RSEs - Sample Data
Description
This record is intended as sample data for a Walkthrough at the UK RSE 2022 Conference titled "MLOps for RSEs". The code that uses this sample dataset can be found on GitHub. There are 2 parts based on the sample problems in the walkthrough
- Classifying wind rotor events
- Clustering weather regimes
Rotors Dataset
This dataset is intended as a machine learning dataset, to train a model to predict the occurrence of turbulent wind gusts called "rotors". These are wind gusts happening on the leeward side of mountains. When they occur near an airfield, this can be hazardous from aviation operations. This data is intended to be used with the code on the Met Office Data Science Community of Practice GitHub repository.
Files:
- 2021_met_office_aviation_rotors.csv - Raw dataset
- 2021_met_office_aviation_rotors_preprocessed.csv - Preprocessed dataset ready for machine learning
- rotors_catalog.yml - Intake Catalog file for the rotors dataset.
More Information:
- MO Data Science Community of Practice GitHub - https://github.com/MetOffice/data_science_cop/tree/master/challenges/2021_falklands_rotors
- Met Office Youtube - What are rotors? https://www.youtube.com/watch?v=jgSZG9SqN_s
- What are Lee Waves? https://www.metoffice.gov.uk/weather/learn-about/weather/types-of-weather/wind/lee-waves\
Weather Regime Clustering
This dataset is a UK and North Atlantic cutout of the Mean Sea-level Pressure (MSLP) field in ERA5 reanalysis dataset produced by ECMWF. This dataset is used for demonstrating an unsupervised learning pipeline.
Files:
- era5_mslp_UK_2017_2020.nc - Gridded dataset of ERA5 reanalysis data.