American Airlines wanted to improve the operation of a specific fleet and tried using analytics to predict the impact of schedule design—including routes, schedule and downtime—on the fleet’s performance. The carrier wanted to start applying “advanced analytics in non-traditional ways to start to better understand the ‘whys’ of operational performance,” says Zachary Shapiro, American’s director of operations reliability.
The unidentified fleet in question moves across an intercontinental spiderweb of hubs, and its on-time departures and completion factors were lower than for other fleets. Ferry flights per scheduled departure were 125% higher. American figured that if it could identify the risks earlier in network planning, it could avoid the ripple effects that cause disruptions, says Shapiro.
In a presentation at GE’s Minds+Machines in October, he said the airline usually addresses reliability issues by examining technical terms, ATA chapters, training, tail swaps and similar inputs—but what about looking at structural factors as well?
“Instead of focusing on technical issues, we widened the aperture to look at other factors,” says Shapiro.
Examining many variables with the help of a GE Aviation data analytics team, American pinpointed various schedule design elements—including the embedded time between shop visits—that affected reliability, says Shapiro.
“You can maximize revenue by lowering ground time, but at what risk?” asks Alex Narkaj, GE Aviation’s director of data and analytics, who worked with the American team.
Processes and behavior options at American and its industry partners, including GE Engine Services, were part of the reliability project’s scope.
“The Line of Flying project evolved our approach from anecdote to analytics,” says Shapiro, adding that American can use the project methodology on other fleet investigations.
“We feel like we are making better decisions from a holistic view of the airline,” and it has opened eyes to advanced analytics, says Shapiro. “We can now look beyond philosophical and gut instinct to evaluate operational decisions.”
Another benefit to the broader, holistic approach is that “this has catalyzed cross-functional collaboration,” he says.
After the initial fleet examination, the reliability team went on a road show across the airline to demonstrate how the analytics project worked and how collaboration was paramount.
“Because so much capability is decentralized across the airline, how do you manage that? We are making an ardent effort to be more cohesive and to leverage each other’s work,” says Shapiro.