An exploratory analysis for predicting passenger satisfaction at global hub airports using logistic model trees

On-time performance during travel is a key factor in determining passenger convenience and satisfaction. On time flight arrivals and departures are dependent on various factors including airport characteristics like the number of flights handled, ground handler's efficiency, disruptions caused by weather conditions, security alerts, queue times at immigration and air traffic congestion. There is a direct causation based relationship on delays, cancellations, re-schedule at the airport with passenger complaints and churn. Airports and airlines can infer signals of passenger satisfaction with the relevant events associated with on-time performance and delays. Hub airports deal with high passenger traffic, connections and operational complexities hence they are good candidates for passenger service studies. In this paper, we had collected datasets for on-time performance, flights for 48 global hub airports and passenger reviews about queue time. Further, we applied Logistic Model Trees (LMT) machine learning method for predicting the level of passenger satisfaction based on factors like an airport on time performance, the number of flights, on-time ranking, average delays and queue time. The results are presented and discussed for further insights and studies.