Predictive Analysis of Platform-Based Tourism: A Time Series Forecast of Monthly Occupancy In Puerto Peñasco, Mexico

Accurate occupancy forecasting is a critical component for planning and decision-making in tourism destinations with a strong presence of platform-based accom-modation. In coastal destinations, demand patterns are typically characterized by pronounced seasonality, which increases uncertainty for both private operators and public authorities. This study aims to identify the most suitable forecasting method for predicting monthly occupancy rates in Puerto Peñasco, Mexico, using historical time series data from the AirDNA platform covering the period from June 2022 to May 2025. A quantitative approach based on time series analysis was adopted. The empirical strategy included exploratory analysis, STL decom-position, and the estimation and comparison of exponential smoothing models (ETS), STL–ETS hybrid approaches, and seasonal ARIMA models. Model se-lection was based on information criteria, forecast accuracy measures, and residu-al diagnostics. The results indicate that the occupancy series exhibits strong and recurrent annual seasonality, accompanied by a smooth long-term trend and mod-erate irregular fluctuations. Under these conditions, the ETS (A, N, A) model provided the most parsimonious and accurate representation of the data. The find-ings highlight the relevance of seasonality-driven forecasting models for plat-form-based tourism markets and provide empirical evidence to support their use as decision-support tools for operational planning and destination management.