Abstract
The tourism industry is an essential part of the global economy, which has shown resilience even during challenging times. Tourism activities have become crucial for individuals, leading to the growth of the leisure sector as an important economic industry that warrants significant policy attention. Therefore, analyses of tourism expenses are necessary for theoretical advancements and practical improvements, considering the varying impacts of the determining factors. While domestic tourism is significant in many nations, measuring its overall economic impact requires evaluating tourism expenditure and consumption. However, studies on the determinants of tourist expenses within national borders are limited. This paper emphasizes the use of microdata for economic and social analyses by employing the “Trips and Holidays” survey conducted by the Italian National Statistics Institute. This survey provides detailed information on vacations within Italy and abroad, including the amount and characteristics of expenditures, travel reasons, and trip durations. This study aims to investigate the influence of primary determinants on average spending in Italy using machine learning’s random forest methodology. By using an individual-level dataset, this study identifies variable effects among selected sociodemographic indicators with significant policy and theoretical implications. The findings highlight that the mean expenditure for domestic tourists in Italy is influenced by journey length, travel party size, and tourist origin and destination regions. Furthermore, the assessment of model predictability demonstrates that the random forest technique exhibits remarkably high predictive capabilities.