CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Forecasting Air Passenger Demand at Palma de Mallorca Airport: A Comparative Analysis of SARIMAX and Econometric Models across Domestic and International Routes

Estudiants que han llegit aquest projecte:

Director/a: PONS PRATS, JORDI

Departament: FIS

Títol: Forecasting Air Passenger Demand at Palma de Mallorca Airport: A Comparative Analysis of SARIMAX and Econometric Models across Domestic and International Routes

Data inici oferta: 26-01-2026     Data finalització oferta: 26-09-2026


Estudis d'assignació del projecte:
    GR ENG SIST AEROESP
Tipus: Individual
 
Lloc de realització: EETAC
 
Segon director/a (UPC): KULJANIN, JOVANA
 
Paraules clau:
Air passenger demand forecasting; SARIMAX; Econometric modelling; Ordinary Least Squares (OLS); Time series analysis; Seasonality; Gravity model; Tourism demand; Route-level forecasting; Palma de Mallorca Airport (PMI)
 
Descripció del contingut i pla d'activitats:
El projecte té com a objectiu l'estudo dels diferents models de proviso de demanda existents. Analitzar la seva aplicabilitat, i la informação tan que necessite com que proveeixen. L'objectiu final és poder aplicar aqueças models a l'aeroport de Palma de Mallorca, per poder estimar el número de passatgers, Operacions i flota a mitra i largo termini.

Pla de treball:
- analisis dels model de demanda
- analisis de l'aplicabilitat dels models. Identificación de limitaciones i avantatges.
- analisis de les condiciones operatives actuals de l'aeroport de PMI. Estudi de les condiciones de contorn de l'aeroport; situació geogràfica, dades macro-econòmiques, integració ne la Xarxa de transport Aeri i multimodal
- aplicació i estimado de la demanda futura a PMI
 
Overview (resum en anglès):
Accurate forecasting of air passenger demand is essential for airlines and airports to plan capacity, schedules and infrastructure, particularly at highly seasonal airports such as Palma de Mallorca (PMI), where summer traffic can exceed four times the winter trough. This thesis develops and compares two forecasting approaches, a Seasonal Autoregressive Integrated Moving Average with exogenous regressors (SARIMAX) model and an econometric regression model estimated by Ordinary Least Squares (OLS), applied independently to five PMI routes: two domestic connections with stable, mixed-purpose demand (Madrid and Barcelona) and three international leisure routes (London, Frankfurt and Paris-Orly). While SARIMAX relies only on the autoregressive and seasonal structure of historical passenger data, the econometric model incorporates explanatory variables such as a gravity-based measure of origin economic mass, tourism occupancy and average length of stay, the Brent crude oil price, a post-COVID recovery dummy and, for the London route, the GBP/EUR exchange rate. It allows forecasting to be combined with causal interpretation of demand drivers. Using monthly Aena traffic data from January 2022 to March 2026 (51 observations), both models are evaluated out-of-sample over a 12-month test window (April 2025-March 2026) using MAPE, RMSE and R², and endogeneity of the occupancy variable is tested via Wu-Hausman. Results show that SARIMAX outperforms the econometric model on four of the five routes, confirming its strength where seasonality is stable and recurring. The exception is PMI-Orly, whose irregular seasonal pattern make the econometric model's explanatory variables decisive, yielding lower forecast error. These findings support a hybrid, route-specific forecasting strategy for PMI, using SARIMAX as the default tool complemented by the econometric model on routes where seasonality is unstable or economic and tourism drivers materially shape demand.

© CBLTIC Campus del Baix Llobregat - UPC