Modeling Airplane Passenger Volatility during the COVID-19 Crisis: a SARIMA and Intervention Analysis

The COVID-19 pandemic in early 2020 had a severe impact on air traffic at Indonesia’s Soekarno-Hatta International Airport, which is among the busiest in the world. This caused a sharp decline in passenger numbers in April 2020, resulting in significant data fluctuations that required statistical intervention. Therefore, to forecast passenger numbers during these fluctuating trends, this study used the SARIMA and Step Function Intervention analysis. The results showed that the Step Function Intervention model was more accurate than SARIMA in predicting the number of passengers at the domestic departure terminal. Based on data for the period between January 2006 and June 2024, the step function intervention model produced MAD, MSE, RMSE, and MAPE values that are smaller than the SARIMA model. The best model, SARIMA Intervention (2,1,0)(1,1,1)12 b = 0, s = 3, r = 1, fulfilled the white noise and normality assumptions with a MAD accuracy value of 82381.85 and MAPE of 9.62%. In addition, the Step Function Intervention Method further reduced the MAPE value by up to 6.93%.