Accurate forecasts are primordial to allow cost-efficient operations while guaranteeing social distance in the pandemic. Uncertainty over the evolution of the pandemic and related travel restrictions though resulted in unpredictable and unstable flight scheduling by airlines. With flight schedules as the basis for forecasting, predictions and planning became incredibly difficult for airports.
With the machine learning models for the forecasting of passengers and bags, developed under the AOP programme, Brussels Airport was able to develop a means to still reach demand forecasts with an accuracy of 86%. This enables the airport to create reference cases and scenarios as a data-driven basis for the development of the recovery plan.