Airport case study

Predict passengers flow to optimize staffing.


The client wants to be able to accurately predict departing passenger flows through the airport, to optimise security control staffing and in turn minimize passenger wait times. Short term spikes in passenger traffic requires the reliance on expensive temporary resources. Recruiting permanent staff is time consuming (vetting, training, onboarding). Solving the issue requires three levels of forecasting:

  1. Week to week: Being able to plan weekly staffing requirements , one month out, on a rolling weekly basis. This allows for the staffing planning, 4 weeks out, based on existing staffing levels, while taking into account planned vacations and other absences.
  2. Intraday: being able to forecast passenger numbers accurately in 15-minute increments throughout the day, to optimize the deployment of available resources at the time they are most needed. This ensures the smoothest possible flow of traffic through the airport throughout theĀ  day, with the resources available.
  3. Month to month: Being able to plan for monthly staffing needs 6 months out on a rolling monthly basis. This allows for the staffing and recruitment planning and the reduction of the reliance on temporary workforce.


Predictive Layer gathers all of the following data points, with the relevant sampling intervals:

  • All the internal HR data from the airport, which gives us staffing levels, planned absences, etc.
  • All of the flight information from all of the airlines
  • All of the passenger information on all of the flights
  • Any short-term or real-time data around weather, flight delays, etc.

With this data, Genius Forecaster was able to begin the training process on all 3 forecasting horizons.

Once trained, the solution was ready to be tested by the client in parallel with benchmarking against passenger data (in other words: validate the short term PL forecast against actual passenger numbers). Having successfully passed these tests, the Predictive Layer solution was deployed into production across all three forecasting scenarios. Genius Forecaster benchmarks show a gain of 20% in precision on passenger flow with 90% accuracy on departure flight filling rate. This real time planning reduces staffing cost by 15%.

Trusted by leading global brands

Shneider Electric
Arcelor Mittal
General Electric