Modelling Buyer Behavior for Airline Bookings

• Goal: Model buyer behavior for airline customer bookings to improve prediction accuracy despite imbalanced target classes.


• Tools/Tech: Python, Pandas, scikit-learn, Random Forest, Sampling techniques (oversampling, undersampling)

This project was part of the British Airways data science certification in which I worked to model buyer behavior for customer bookings using random forest methods. The main thing I learned from this project was how the balance of target variable classes can affect the performance of a model, and how to use techniques such as oversampling and undersampling to mitigate this.