Space-time correlation analysis of quality-of-service at major US airports-O´Hare International and Minneapolis Airports: a case study

In this paper we study and analyze the space-time correlation among different variables in the NAS such as weather, capacity, delay, and cancellation at hub airports. A Quality-of-Service (QoS) Index is formalized as a vector of delay, cancellation and variability of the trade-off relationship between these metrics. A user-friendly graphic tool was first developed in C++/OpenGL to visualize geometrically and chronologically the large, complex data sets of ASPM and BTS. This tool can show how the QoS indexes evolve at major airports visually as time and weather change and help determine their relationships. We next explored the BTS dataset to study the space-time correlations of QoS Indexes for one particular case study involving ORD and MSP airports whose operational correlation can be easily identified by the tool. The analysis was done at the airline level, and revealed patterns of behavior that are airline-specific. Delay and cancellation time series were aggregated using a weighting function for cancellation. Further investigation on Northwest Airlines´ aggregated time series showed a significant positive cross-correlation between the airline´s late arrivals at ORD and those from ORD to MSP after a time lag of 135 minutes.