Reputation assessment and visitor arrival forecasts for data driven tourism attractions assessment

Tourism is vital for most historical and cultural cities. In the context of Smart Cities, there are numerous data sources in tourism domain that could be analyzed to monitor and forecast a range of different indicators related to touristic locations and attractions. In this paper, we propose a framework which exploits social media and big data to forecast both online reputation and touristic attraction presences. To this end, some techniques have been tested and proposed on the basis of machine learning, deep learning, causality assessment and explainable Artificial Intelligence, so as to provide evidence of the relevant variables for each prediction and estimation. An approach has been introduced to analyze the explainability of the proposed solutions, i.e., a multilingual sentiment analysis tool for social media data based on transformers to compare data sources as Trip Advisor and Twitter. Furthermore, causality analysis has been performed to evaluate the temporal impact of social media posts and other factors with respect to the number of presences. The work has been developed in the context of Herit-Data, a European Commission funded project on the exploitation of big data for tourism management and based on the Snap4City infrastructure and platform. Herit-Data has developed solutions for 6 major European touristic locations. In this paper, some of the solutions developed for Florence, Italy and Pont du Gard, France, are reported.