MONTUR project: Dataset for understanding and forecasting tourist flows.

Area of Science:

• * Data Science and Artificial Intelligence
• * Transportation and Tourism Management

Background:

• * Effective management of tourist flows is crucial for regional economic and social policies.
• * Existing methods often lack real-time data for dynamic resource allocation.

Purpose of the Study:

• * To develop and validate an advanced system for monitoring and forecasting tourist flows in the Aosta Valley.
• * To leverage distributed sensor technologies, cameras, and machine learning for real-time traffic and presence data.
• * To enhance decision-making for regional tourism resource management and operational efficiency.

Main Methods:

• * Integration and analysis of over 41 million vehicle passages from traffic detection portals.
• * Computational optimization of vehicle flow data, focusing on checkpoints and categories.
• * Application of the eXtreme Gradient Boosting (XGBoost) algorithm for time series forecasting due to high data stationarity.

Main Results:

• * The eXtreme Gradient Boosting (XGBoost) algorithm demonstrated superior forecasting accuracy compared to Deep Learning and other Machine Learning models.
• * Optimized data processing reduced dataset size while maintaining analytical integrity.
• * The system provides real-time data for improved traffic and tourism resource management.

Conclusions:

• * The developed system significantly advances the capability to monitor and forecast tourist flows.
• * Accurate forecasting enables better management of resources, traffic, and visitor experiences in the Aosta Valley.
• * The study highlights the effectiveness of XGBoost for stationary time series forecasting in tourism applications.