Simulation of recommender systems driven tourism promotion campaigns

With overtourism becoming an increasingly widespread problem, it is becoming more and more important to better analyse trends of tourists’ arrivals in a region, and to foresee the effects that alternative promotion campaigns, if delivered on an online destination platform, may have on the distribution of tourists in the region’s destinations. To facilitate this analysis, the study proposes a tool that enables a Destination Management Organization to simulate the effect of an online promotion campaign, which uses Recommender Systems techniques, to select which destinations are promoted to each tourist. A case study is developed in South Tyrol, a highly-visited province in the Italian Alps. In the simulated scenario, tourists, who visited the region in the past, are simulated to be exposed to promoted destinations. Each simulated tourist can choose an option among the tourist’s actual choice and other destinations that are promoted. The tool allows setting up simulations, and visualising their results. It also offers data analysis functionality to inspect tourism arrivals data. The experiments carried out in the study reveal a number of important effects on the expected distribution of tourists, and helps identifying which promotion campaign is likely to improve the sustainability of tourism in the province. The tourism data analysis and promotion campaign effect simulation tool developed in this work was positively evaluated by tourism domain specialists and usability survey participants.