Wang, ShenShenWangMacNamee, BrianBrianMacNamee2017-12-142017-12-142017 IEE2017-09-17http://hdl.handle.net/10197/9113International Smart Cities Conference (ISC2), Wuxi, China, 14 Sep - 17 Sep 2017AbstractAn increasing number of people use smartphoneapplications to plan their trips. Unfortunately, for variousreasons, bus trips suggested by such applications are not asreliable as other trip types (e.g. by car, on foot, or by bicycle),which can result in excessive waiting time, or even the needto revise a planned trip. Traditional punctuality-based busservice reliability metrics do not capture route deviations, whichare especially frequent in rapid changing urban environmentsdue to rapidly changing road conditions caused by trafficcongestion, road maintenance, etc. The prevalence of GPS dataallows buses to be tracked and route deviations to be captured.We use such data to propose and calculate a novel reliabilityscore for bus trips. This score is a linear weighted combinationof distance, time, and speed deviations from an expected, predefinedbus trip. GPS trajectory data is large and noisy whichmakes it challenging to process. This paper also presents anefficient framework that can de-noise and semantically splitraw GPS data by pre-defined bus trips in citywide. Finally,the paper presents a comparative case study that applies theproposed reliability score to publicly available open bus datafrom Rio de Janeiro in Brazil and Dublin in Ireland.en© 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other worksMachine learningStatisticsConference Publication10.1109/ISC2.2017.80908432017-12-04https://creativecommons.org/licenses/by-nc-nd/3.0/ie/