Internet as a Sensor: a Case Study with Special Events

This paper is about utilizing the web as a predictor for transport demand. The quality and quantity of information on the web, updated frequently, turns it into an enormous sensor of present occurrences and into an oracle for future occurrences. At the same time, the quality of data coming from public transit, taxi and private car usage is increasing and its availability starts to become a fact in many cities in the world. Such data revolution will impact transport demand modeling, particularly making elaborate approaches feasible for real time applications. The authors present here their approach to this challenge and provide particular focus to a case study with special events in the city-state of Singapore. The authors build an origin/destination trip prediction model that uses information extracted from the Internet together with tap-in/tap-out OD data from public transit (EZLink card).