TRB 2016 Blue Ribbon Committee
16th National Transportation Planning Applications Conference

An Exploration Of ‘Big Data’ Sources To Inform Best Practice Travel Time Studies - Lessons Learned From Metro Vancouver

Corresponding Author: Mohamed Mahmoud, TransLink

Presented By: Fearghal King , TransLink


There has always been a need for Transportation Authorities to understand and communicate travel patterns and evaluate system performance within their respective regions. Such insight is a critical component of the strategic planning, development, and operations of transportation services and infrastructure. Furthermore, this insight helps justify capital investments, and provides an opportunity to link the policy change or investment to systems’ performance.


Traditionally, travel time information was obtained by conducting costly and time consuming travel time surveys using GPS equipped probe vehicles. Such surveys were limited to relatively low spatial and temporal resolutions. Throughout the past decade, there has been a rapid increase in the existence and usage of mobile/cellular technologies and in-vehicle navigation systems which often collect traffic data passively (e.g. Google maps, INRIX, TomTom, etc.). These new capabilities provide a significant opportunity for transportation agencies to analyse travel patterns across space and time with unprecedented granularity in a cost efficient manner. However, each of these technologies comes with its own limitations and the provided datasets are subject to several sources of bias. That is, dealing with this type of ‘big data’ requires careful investigation and validation prior to drawing conclusions on travel patterns.


This presentation explores how new datasets, while still in their infancy, are paving the way towards understanding automobile travel times and associated congestion in Metro Vancouver. Specifically, the presentation draws upon travel time data explored/obtained from different sources to help develop relevant congestion-related metrics and the shortcomings with each data source. We then proceed with the analysis using one primary data source – Google Maps API, where the process of data acquisition, mapping, and aggregation is outlined. Travel times are analysed across three spatial layers: (i) sub-centres (OD level), (ii) sub-regions (municipality level); and (iii) the road network (segment level). The data passes a comprehensive data validation test, ensuring confidence in the source. This work paves the way for transportation agencies to exploit similar datasets in order to develop robust and meaningful travel time indices on an on-going basis through a scalable process. We summarise with lessons learned dealing with various datasets and conclude with recommendations on the appropriateness of using passive data for travel time studies.


Discuss This Abstract