Estimation of travel time from GPS data

Abstract

In this “big data” era, a huge amount of data is collected on a daily basis, from the logistics to recreational sector and from customer shopping history to trains operation logs. However, these raw data are not being utilised enough to provide insights for better planning and operations. In the case of the taxi industry in Singapore whose total taxi fleet size is close to 25,000 with a daily total of 600,000 trips. This could be a potentially good source to give us insights about the traffic conditions, travel patterns etc. In this thesis, simple solution to the travel time estimation problem using nearest-neighbour methods is presented. The proposed methods have a 5% increase in estimation accuracy at 0.22 from a baseline method, while they suffer from high time complexities.