Predictive patterns in aviation delay management

In a rapidly evolving world, airlines have to continuously adapt and face new challenges to an ever-changing web of factors that will affect their methods of informed decisions. This project aims to analyze the aviation industry's approach on its delay management to then develop a framework of predicting and quantifying airline delay, which will be done around the perspective of the airport of Madrid-Barajas through an amalgamation of data sourcing and data science techniques. The study unfolds in several critical phases, beginning with the definitions of the current delay management in the aviation industry and demonstrating which are the most significant for this study. Secondly, the project will rely on different types of datasets around Madrid's airport, including, real historical flight data for encompassing departures/arrivals, and airline specifics. Additionally, a significant emphasis is placed on incorporating weather data, with a keen focus on wind gusts and wind directions. These and more variables will be cleaned, manipulated, and be considered in different ways to see how if may affect the delay in the next step of the project. Subsequent to data collection, the project transitions into the data analysis phase, which will consist of extracting key variables that will be used for the prediction phase of the project. Making sure what variables to use is essential as this will shape how the prediction models are going to use them and to what level we can draw conclusions. The last phase simulates a present-time situation to predict the delay throughout a given day after a particular hour, trying to predict the following hours of activity, which will give us insights on how it can be managed if whether it is recuperated or enhanced. This will be done by exploring two machine learning models: the Random Forest Regressor and the XGBoost model, which both serve purposes in different ways, allowing for a multi-modal analysis point of view. Overall, this project