Use of Mobile Phone Data in Regional Traffic Models: Experience from Highways England
Siamak Khorgami, AECOM, Graham Powell, Highways England, Reza Tolouei, AECOM
The paper considers the strengths of trip matrices estimated from mobile phone data compared to conventional sources. A framework is designed and applied to verify the data before developing prior trip matrices.
Authors: Siamak Khorgami, Graham Powell, Reza Tolouei, Paul Hanson
Use of new sources of big data, such as mobile phone data, provide opportunities to better understand travel patterns across wide areas, but use of such data presents its own challenges in order to ensure the models are fit for the purposes for which they are intended.
Highways England commissioned five Regional Traffic Models (RTM) covering the whole of England to provide a consistent evidence base to develop schemes defined in the Road Investment Strategy. The development of the RTM trip matrices has drawn heavily on mobile phone data. The RTMs have involved five different study teams across a number of companies and locations, so the approach adopted has been one of collaboration, to a level unprecedented in the industry.
The use of mobile phone data in trip matrix development is a new technology and there are no existing guidelines or established practice; this represents a significant risk to a project of this scale. Highways England established Technical Consistency Groups, including the matrix development group, that span across the regional models to ensure that all elements of the work are undertaken in a consistent manner and that best practice is shared and developed across the RTMs. It is hoped that the lessons learnt can assist development of transport models elsewhere.
This paper sets out the key considerations for matrix development and the lessons learned regarding its use. The paper discusses the strengths of trip matrices estimated from mobile phone compared to conventional sources of information. It also discusses and demonstrates the key weaknesses and uncertainties associated with trip matrices derived from mobile phone data, as the data are not collected exclusively for the purpose of transport modelling. The processes used to verify and adjust the mobile data are set out together with the rationale as to why each is necessary.
Association for European Transport