Baggage handling operations in airports have become substantially more challenging due to the surge in air transportation demand witnessed in the last three decades combined with an even higher pressure on the operations following the COVID-19 pandemic. The intricate nature of checked baggage requirements, impacting resource allocation and personnel scheduling, necessitates an integrated approach for problem-solving. This research paper aims to enhance baggage handling operations by predicting the baggage factor (BF) for individual outbound flights. The BF represents the ratio of checked baggage items to the number of passengers aboard an aircraft. The objective of this study is to create a forecast model that predicts baggage factors for individual outbound flights over a time span of 7, 30, and 60 days, and subsequently construct a baseline model that leverages the forecasting outputs to optimise baggage handling processes and allocate resources effectively. A novel approach is proposed that employs historical flight data within gradient boosting models to forecast the baggage factor for future flights. Additionally, a case study is conducted by developing a Mixed Integer Linear Programming (MILP) model to minimise space utilisation within a baggage handling facility during the busiest period of the day, employing as few Make-Up Areas (MUAs) as possible. The results indicate that LightGBM, a gradient boosting technique, outperforms other gradient boosting techniques in terms of performance and computation time, achieving an accuracy score for the BF prediction ranging between 78-83% for the three forecast periods. Leveraging these predictions, the MILP model demonstrates that only 3 to 5 MUAs are required in an ideal situation in the baggage handling facility during the busiest period on various days.