A utility-maximization model for retrieving users' willingness to travel for participating in activities from big-data

Dense cities with complex transport infrastructure and numerous Places of Interest (POIs) pose challenges to interpersonal interactions since the location and the time of conducting a joint activity are rarely agreed unanimously from all agents, mainly due to lack of information of each agent's schedule and activity preferences. In this paper, a utility maximization model for capturing automatically users' willingness to travel a certain distance for participating in different activity types based on user-generated data is introduced. The model is based on Big-Data analytics on user-generated data from Social Media that can offer valuable insights into users' preferences and improve our understanding on their decision-making mechanisms for selecting joint activities. Unlike static approaches, the utility model of each agent is developed to incorporate continuously updated data feeds regarding agents' activities (i.e., timing of visiting a POI, type of POI). The agent-utility model is implemented and tested via utilizing geo-tagged, social media data from 65 individuals from the dense city of London (crawling campaign from November 2012-January 2014). After implementing the utility-maximization model, all users are clustered based on their willingness to travel similar distances to participate in certain types of activities, facilitating the development of applications for suggesting joint activities among agents with similar profiles.