View notebook (NBViewer): https://nbviewer.org/github/IngaPosiunaite/taps-aff-weather-ml/blob/main/taps_aff_weather_ml.ipynb
“Taps aff” is a Scottish phrase meaning “tops off”, commonly used when the weather is unusually warm.
This project uses machine learning to predict whether a day in Glasgow is a “taps aff” day based on historical weather data.
The project applies regression and binary classification to estimate missing weather values and predict warm days between 2023 and 2025. It follows a complete machine learning workflow, from data preparation to model evaluation.
- Predict missing weather values using a regression model
- Engineer date features to capture seasonal patterns
- Merge weather data with labeled “taps aff” records
- Train a binary classification model to predict warm days
- Evaluate model performance using test data
- Handling missing data with regression
- Feature engineering for time-based data
- Training and evaluating classification models
- Working with real-world weather datasets
Python, Pandas, NumPy, Matplotlib, TensorFlow/Keras, Scikit-learn