Validating tourism modeling with real data

The framework recently proposed by Nieto-García et al. (2023) presents a provocative critique of tourism research, identifying "researcher myopia" as a systematic failure to measure actual behavior. This analysis draws empirical support from the MONTUR project (Alderighi et al., 2025) (Real-time monitoring and forecasting of tourist flows in Aosta Valley), a multi-year behavioral monitoring initiative that collected optical vehicle detection data from 14 sensor portals across major transportation corridors in the Aosta Valley region of Italy. The system recorded 41 million individual vehicle passages over three calendar years (January 2021–December 2023), classifying each passage into seven behavioral categories based on observed visit frequency patterns. Following the methodological checklist for tourism researchers proposed by Nieto-García et al. (2023), this analysis addresses their call to operationalize behavioral measurement by: (1) defining observable outcomes rather than stated intentions as the dependent variable; (2) employing validated sensor technology for continuous measurement; and (3) comparing multiple analytical approaches to identify optimal forecasting methods. Machine learning models applied to this dataset achieved quantifiable forecasting accuracy that can be compared across algorithms and validated against subsequent observed behavior. Tourism professionals often encounter a troubling pattern: models that demonstrate strong theoretical foundations and statistical validation in controlled settings often underperform when confronted with real-world complexity, seasonality, disruptions, and data heterogeneity. The problem is not that theory is weak, but rather that the pathway from elegant theoretical mechanisms to reliable forecasting tools remains underdeveloped. This commentary is directly informed by the conceptual framework advanced by Nieto-García et al. (2023). Their critique of "researcher myopia" had a concrete methodological impact on the present study by motivating the exclusive use of observed behavioral data rather than stated intentions. In line with their recommendations, this paper operationalizes tourism behavior through large-scale, sensor-based monitoring, and validates forecasting models against real tourist flows. In this sense, the present analysis represents an empirical instantiation of the impact of Nieto-García et al. (2023), demonstrating how their conceptual contribution translated into research design choices, collaboration with destination authorities, and policy-relevant evidence grounded in actual tourist behavior.