The Edinburgh Earth Observatory (EEO) AGI-Scotland seminar series were delighted to host Stephen Cragg, transport planner for Transport Scotland on Friday the 31st of March, 2017. Stephen’s talk formed part of the seminar series’ ‘Future’ theme, in which GI professionals across a range of fields and disciplines examine the pervasive nature of technology, and the ways in which it is changing our world for the better (or for the worse).
The seminar was intriguingly entitled ‘Future Transport: I know where you were last night!”. Steven stated that the title was, in itself, a technique to garner curiosity from among the audience, and a question which he sought to answer in the concluding remarks of his talk.
Steven amusingly began his discussions on the premise that he knew his left from his right, but not necessarily his east from his west, setting the scene for a spatially-minded audience. He followed-up by outlining in turn, the range of high profile transport projects in which he has been involved in, including The Queensferry Crossing, the completion of the M74 south of Glasgow (near Auchenshuggle Bridge no less), Edinburgh-Glasgow rail improvements, and the dualling of the A9.
He noted that his involvement in such projects were at the earliest stages of development, the one who posed the question: ‘do we really need this? What is the business case?’ With such an emphasis on the business case, Stephen argued that he needed to understand who was going to use such large-scale transport projects, why would they use it, and crucially where were they going from and where were they going to?
With this context in mind, Steven turned his attention to how such information could be uncovered. He spoke of his frustrations with the transport-limited nature of the questions posed in the census, with the only gauge being the address of the participant’s workplace for which he/she travels to. He then turned his attention to the Scottish Household Survey through which participants were asked to provide a travel diary, but he noted its limitations in scope, with only 15,000 annual participants extrapolated to a population of over 5 million. He estimated that a discussion with a driver costs Transport Scotland around £10, and noted that there are around 4 billion personal journeys in Scotland annually, ultimately totally a rather expensive endeavour.
With such expense associated with traditional surveying, Steven then posed the question: “can I tap into a whole range of remote monitoring systems?” By remote monitoring systems, Steven was referring to Automatic Number Plate Recognition (ANPR), Bluetooth Detection (BD), and Mobile Phone Tracking (MPT).
Firstly, ANPR was discussed in relation to the collection of journey time data, their “raison d’etre”. Steven noted that while ANPR sensors were not installed for his purposes, he wondered could he tap into their data, and commandeer said data for his benefit. Secondly, BD installed by Transport Scotland, again for journey time data, could he tap into this data? Finally, MPT based on the context of your mobile phone knowing where it is, and by extension knowing where you are. Steven did not believe that your mobile phone could be used to track you to the finest level of detail, but instead within a larger ‘cell’ as governed by a mobile phone tower’s coverage.
Steven then turned his attention to the common challenges associated with accessing data for purposes beyond which it was originally collected for. Steven found it surprising that he was unable to simply access data from speed cameras as a transport planner for Transport Scotland, and noted that Home Office-style approval would be required for such access. In contrast, mobile phone data is becoming increasingly available in terms of the selling of ‘movement data’. Steven then sparked an interesting discussion with regards to the ways in which we can translate number plate data, speed data, and movement data, into people data, more personalised to the individual(s) concerned.
The inherent bias prevalent in such data was discussed in terms of those people who have access to modern technology and those that do not. For example, the data would be skewed to those people who own a modern smartphone, but how do we account for those people that do not?
Steven then asked “what does location actually mean?” Through that, he discussed issues of geographical scale, so how precise does the location have to be? Street-level data, neighbourhood data, city data, national data, and so on? This question becomes more acute when considering the urban-rural divide, and the issues of data coverage. This related to the disentanglement of data in the context of assessing a ‘stop’ in a journey from A to B. Was this ‘break’ in movement simply due to lack of adequate sensor coverage, or was it a genuine detour, and if so why, where, and for how long?
Steven then turned his attention to the ‘bread and butter’ of a transport planner, journey purpose. He argued that this was key due to the fact that the responses of travellers to the decisions and interventions taken by the likes of himself, depended upon the purpose of the journey they were making. He spoke of longitudinal monitoring as a means to determine journey purpose in which repeated journeys recorded via a smart device could provide indications of places of work, study, residence, and so on.
Drawing to a close, Steven considered issues related to privacy and the ethics of third-party data selling. He interestingly asked the audience if they regarded a car registration plate to be personal, private data, to which a mixed response was provided by the audience. The key issue of data linkage was Steven’s final point, with the collected data being aggregated with other data sources to construct a movement profile of the individual. For that reason, Steven argued that a car registration plate could indeed be considered personal, private data.
Steven ended his thought-provoking seminar arguing that while he doesn’t know where I was as an individual last night, he could know where we were as an audience last night on a larger, aggregate level.
Blair JH Bell
MSc. Geographical Information Science
The University of Edinburgh