This is the repository for "Data-driven safety indicators for flight operations around Schiphol Airport". Air traffic management is one of the most complex systems that humans have ever created. In this system, ensuring safe aircraft operations has the utmost importance. The possibility to collect large amounts of data and the availability of data mining techniques allow new ways to monitor air traffic. This thesis proposes an innovative proactive risk management strategy that uses safety indicators.
These indicators combine ADS-B data and data mining techniques to identify anomalous safety events and their precursors. ADS-B technology allow to collect vast quantity of operational data and it includes information about position, velocity, climb rate, heading, and identity of the sender, and it has a latency of 1sec. The data mining techniques include rule-based algorithms, clustering, Support Vector Machines, Gaussian Mixture Model, Neural Network and Principal Component Analysis.
These techniques return statistically significant events, and a post-processing analysis is required to discover truly operational significant occurrences. Schiphol Airport is selected as the case study for the analysis because it is one of the biggest and most complex airports in the world. The objective is to design safety indicators that reveal the underlying safety performance of the airport.