Observing transient behavior during Hurricane Sandy through passively collected data

Abstract

Transients, such as tourists or business travelers, often fly into large cities and subsequently have to rely on public transportation for movement within the city. It is difficult to observe transient behavior in the event of a disaster through traditional survey methods. Thus, transient behavior has not been understood as well as resident behavior, which presents a challenge to emergency management. This study used taxi trip records that are passively collected year-round in New York City (NYC) to observe transient behavior during Hurricane Sandy (2012). The linkage between transients and taxi trips was built upon the fact that many transients choose hotels as accommodation and depend on taxis during their travel in NYC. This study extracted and analyzed taxi trips originating from hotels shortly before and after the hurricane landfall, which cover transient outbound activities in a week. In the pre-landfall period, it was found that transients who evacuated out of Manhattan were more likely to leave hotels about 18–24 h before hurricane landfall; and those who stayed made more shopping trips within Manhattan. In the post-landfall period, transients were more likely to relocate on the second day after hurricane landfall; some transients headed to local airports once limited service was resumed. This is believed to be the first study to use passively collected data to observe transient travel behavior in the event of a hurricane as well as on normal days.