Urban Travel Time Estimation in Greater Maputo Using Mobile Phone Big Data

Abstract

The rapid growth of urban populations may outpace the development of needed urban infrastructure, such as related to transportation, therefore, resulting to inadequacy of public transportation services and traffic congestion. Travel time is a key component to describe traffic efficiency and has always been an important element to study and control. In this paper, we leverage millions of these records in order to estimate urban travel time in Greater Maputo, Mozambique. To validate our results, we compared the estimated travel time with that computed from GPS trajectories for 600 users. Furthermore, we show how the large number of phone records allow us to segregate travel times of three different time categories, weekday rush hour, weekday non-rush hour, and weekends. Unlike mobile data, the low number of GPS points enabled us to only estimate average travel time without distinguishing different time categories. However, we found 87% linear correlation between average travel time estimated from call details records and GPS datasets. This, to some extent, support the potential of mobile data to monitor traffic time and assess the impact of new transportation infrastructure on urban travel time.