Can Dynamic Social Events Impact Our Travel Patterns And Trends?

Abstract

In recent years, travel route planning and recommendation systems have earned a great popularity. Although several works have focused on suggesting lists or sequences of Point of Interests (POIs), they do not address POIs with dynamic spatio-temporal happenings (i.e. live events) that may attract the tourist's attention. In this paper, we present a system aiming to generate dynamic and personalized travel routes at a given destination based on tourist preferences and dynamic events that may attract large number of tourists. To achieve this, a rule-based classification model is developed to classify geotagged posts collected from Twitter into tourism related posts and non-related posts and further classify the tourism-related posts into types of events in terms of promoting the tourism industry such as festivals and sport tournament. Tourism-related events are detected from the classified posts after clustering all posts of each event type based on location, and then determine unusual temporal slices that may contain significant events and meet tourists' interests and preferences. Our system finally generates a travel route of the top ranked attractions (events) by computing the similarity between the properties of each attraction and the tourist's preferences and the attractions, the top ranked attractions are visualized on a dynamic map. Our system is applied on a real dataset collected from Twitter from the city of Dubai in United Arab Emirates.