An Approach to Evaluate the Impact on Travel Time of Bus Network Changes

This paper proposes an approach to evaluate the impact of bus network changes on bus travel time. The approach relies on data obtained from buses equipped with GPS devices, which act as mobile traffic sensors. It involves three main steps: (1) analysis of the bus network to determine which road segments are frequently traversed by buses; (2) computation of bus travel time patterns by segment; (3) evaluation of how much the bus travel time patterns vary when bus network changes take place. The approach combines graph algorithms and geospatial data mining techniques. It can be applied to cities served by a dense bus network, where buses are equipped with active GPS devices that continuously transmit their position. The paper applies the proposed approach to evaluate how bus travel time patterns in the City of Rio de Janeiro were affected by traffic changes implemented mostly for the Rio 2016 Olympic Games.