An Analysis of Airline Passenger Delays Using Flight Operations and Passenger Booking Data

Abstract

The performance metric used to evaluate on-time performance in the US airline industry is flight-based, measuring the number of flight legs with arrival delay of 15 minutes or more. We analyze airline passenger operations and schedule performance and conclude that this flight-based performance metric does not accurately reflect delays to passengers, in part because it does not recognize the long passenger delays resulting from flight leg cancellations and missed connections. Using passenger booking and flight operations data from a major US airline, we develop a Passenger Delay Calculator to compute passenger delays and to establish relationships between passenger delays and cancellation rates, flight leg delay distributions, load factors, and flight schedule design. We conclude that existing flight-based metrics underestimate passenger delays, and hence, using the insights gained in our analysis and publicly available airline schedule data, we define new passenger-centric metrics for schedule reliability. We also investigate the effects on passenger delays of improved on-line optimization decision models to recover from schedule irregularities.