Sage Journals: Discover world-class research

Abstract

This article explores the process by which intelligent transport system technologies have further advanced a data-driven culture in public transport and traffic control. Based on 12 interviews with transport engineers and fieldwork visits to three control rooms, it follows the implementation of Real-Time Passenger Information in Dublin and the various technologies on which it is dependent. It uses the concept of ‘data ratcheting’ to describe how a new data-driven rational order supplants a gradualist, conservative ethos, creating technological dependencies that pressure organisations to take control of their own data and curate accessibility to outside organisations. It is argued that the implementation of Real-Time Passenger Information forms part of a changing landscape of urban technologies as cities move from a phase of opening data silos and expanded communication across departments and with citizens towards one in which new streams of digital data are recognised for their value in stabilising novel forms of city administration.

Keywords

Intelligent transport systems real-time information smart city Big Data organisational change

Introduction

There is a quiet ‘revolution’ underway in the transport sector as intelligent transport systems (ITS) technologies are used to increase efficiencies and integrate urban functions, through an alliance of well-established transport technology companies in partnership with city engineers and technologists. ITS originates in the context of managing and coordinating road use, but interfaces with air, rail and water systems (Williams, 2008: 3). While discursively enveloped in the promise of the smart city in recent marketing programmes such as that of Dublin (Coletta et al., 2018a; Kitchin et al., 2018), the history of ITS extends back further over several decades with the development of CCTV-supplied traffic control rooms, junction controller software, and scheduling systems for public transport. Therefore, it provides a counter-example to the creation of new urban corporate districts branded as ‘smart’ (Wiig, 2019), or reductive concepts of a universal urban science (Shelton, 2017), inasmuch as it accounts for a gradual evolution of ‘smart’ or intelligent technologies within a specific sector.

ITS forms the basis for a longitudinal study on how data-driven organisational control is being enacted in our cities, drawing attention to the concept of ‘data-ratcheting’ to show how these data-driven changes are iterative and leveraged off previous innovations. The evolution of ITS is dependent on long-standing commitments to shared infrastructure, and indicative of a shift to increasing autonomous management of public transport and traffic while overseen by human operators. The article seeks to contribute to the theoretical literature on standards and Big Data by building an explanatory account of data-driven organisational change based on a case study on the deployment of Real-Time Passenger Information (henceforth RTPI) in Dublin. It focuses primarily on the implementation of RTPI on bus services and considers the wide array of technologies on which RTPI is reliant, including locational technologies, telecommunications and information standards.

The following section reviews existing theory on data infrastructures in relation to its bearing for understanding the deployment of new technologies. This is followed by a description of the fieldwork and methods underpinning the empirical content. The subsequent three empirical sections then detail the deployment of RTPI in Dublin, largely in temporal order and identifying the various phases of development. This covers earlier trials, the negotiation of common information standards, forms of data-driven behavioural or procedural change, and the consolidation of operational intelligence technologies. The penultimate section further discusses data ratcheting in relation to procedural and organisational change. The conclusion then discusses the evolution of data-driven change in transport and speculates on further research directions.

Combining studies of the mundane with data assemblages

The study of infrastructure has covered both epic transformations, such as the electrification of Western societies (Hughes, 1993) and the mundane ‘boring things’ (Star, 1999), such as the file folder system for computing and its consequences for systems management (Yates, 1993). In the former, Hughes’ panoramic view of large technical systems repurposes the military term ‘reverse salient’, pertaining to line formation, to define where blockages in one domain may be cleared by progress in another. Infrastructure becomes a means of analysing how systems travel and extend into local contexts, and a basis for forming concepts to explore their many contingencies and the development of common standards of information exchange.

Infrastructure studies blend the archival patience of the historian with the attentive eye of the ethnographer and require a predilection for interrogating technical reports and manuals. In this vein, Edwards (2010) covers the pioneering work of climate observers in the 19th and 20th centuries and the process of reanalysis, where scientists reconcile data from multiple sources into uniform global datasets. This large temporal scale opens up concepts such as ‘infrastructural inversion’, where ‘historical changes frequently ascribed to some spectacular product of an age are frequently more a feature of an infrastructure permitting the development of that product’ (Star and Bowker, 2006: 233). It also facilitates the analysis of positive and negative externalities associated with standards, such as Meinel's historical study (2004) on the contingencies of the field of microbiology on particular visual and mechanical model kits.

The digital revolution has expanded the remit of infrastructure studies. Star and Ruhleder's (2001) seminal study on the creation of a unified, computer-based and networked worm system for scientists examines the frictions that occur around arcane choices of networking technologies or operating system, providing a useful distinction between first-, second- and third-order issues. First-order issues concern direct matter-of-fact phenomena and can be solved with existing resources and processes, such as finding and enabling a software option. Second-order issues reflect unforeseen contextual effects, such as the choice of one suite of procedures or standards over another and the path dependencies that result. Third-order issues are inherently political, involving ethical frameworks, theoretical paradigms and schools of thought. They may arise from combinations of first- and second-order issues, such as where an awareness emerges that legacy technologies are constraining accessibility, opening the path towards more fundamental questions on the collective good.

Following Kitchin (2014a), we can associate the centuries-old Big Data of climate science referred to above with volume and variety. However, it is with ITS and similar forms of sensor-fed, large-scale digital networks and models that we can see the increasing importance of velocity (i.e. real-time or close to it). ITS relies on complex assemblages of information networks, human engineers and controllers, transport policies, road-side sensors, in-vehicle computers and global positioning systems to choreograph in real-time an increasingly wide array of objects in space. It is part of a shift towards data-driven design and maintenance of urban transport systems, drawn into discourse on the ‘smart city’ through including ‘smart travel’ or transport in funding programmes and established practices of classifying smart technologies (Giffinger and Pichler-Milanović, 2007), as well as through the inroads into transport sought by data analytics companies.

Calls have been made to study actually existing ‘smart cities’ (Coletta et al., 2018a; Shelton et al., 2015), and engage with software by analysing critically how data interact with urban space and society (Kitchin and Dodge, 2011). This can take the form of following the data and ‘attending to the sociotechnical fuzziness of data as it falls between epistemological problems, material infrastructure and organizational concerns’ (Coletta et al., 2018b: 6). It implies looking to the ‘proxies’ of data, where data flows are transformed, bifurcated, or collated; whether that be errors and anomalies in sensor networks and transport systems (Reed, 2018; Reed and Webster, 2010), or the rollout of sophisticated city-wide projects in partnership with industry giants (Shapiro, 2018). Furthermore, it requires attention to the power relations and values implied as these data are recruited into new modes of data-driven urban governance.

Shelton (2017) shows how Big Data-inspired visualisations of vacant lots may depoliticise ethnic and social inequalities, normalising the principles by which such inequalities are largely left in place. Consequently, it is useful to account for the totality of politics, data, values and materialities that underwrite such visualisations or policy tools. Kitchin and Lauriault (2014: 1) advance the concept of a ‘data assemblage that encompasses all of the technological, political, social and economic apparatuses and elements that constitutes and frames the generation, circulation and deployment of data’. It functions as a means of expanding discussion to how data reshape society, and resultantly, how data-driven change leads to further developments and the creation of new path dependencies. This aids in drawing attention to the politics of Big Data (Shelton, 2017), the rights of citizens to the digital city (Foth et al., 2015), and data literacy (Gray et al., 2018).

This article takes its cue from data assemblages as a means of exploring how urban datafication reshapes urban management and control. Kitchin (2014b: 25) pursues data assemblages in relation to how data can usher in new regimes of data-driven societal control, tabling a broad array of constituent factors. The framing of RTPI as a data assemblage opens discussion into how ‘each apparatus and their elements frame what is possible, desirable and expected of data’, including governmentalities, materialities, practices and systems of thought. This article is particularly concerned with procedure; how transport systems have integrated data-driven technologies into their internal modes of organisation and coordination rather than their changing relationship to passengers and broader society.

Modern intelligent transport technologies, it is maintained, represent an instance of urban datafication and consolidation, as organisations experiment with Big Data in the interests of increased efficiency and performance. This process of urban datafication can be described with recourse to three translations. Firstly, data expansionism occurs through the deployment of sensing technologies and ICT infrastructure to create new datasets. Concepts from infrastructural studies such as ‘reverse salients’ help explain how data platforms come into being as new technologies facilitate change and transport authorities and providers build necessary infrastructure. Secondly, as new data sources become available, data experimentalism ensues as a range of actors compete to create new services and products. Finally, a third phase of operational consolidation can be delineated, as new forms of data-driven behavioural change, based on the dynamic treatment of real-time data, are encoded into procedures and organisational technologies. A continual negotiation and policing of standards is necessary to ensure data-driven processes can function to a required degree of resilience, for which the study of first-, second-, and third-order issues aid in our comprehension of how this occurs on different levels. Finally, the titular concept of data ratcheting functions as a creative interlacing of these final and overlapping phases as new functions, products and services are discovered and implemented in a context of data-driven rationalisation.

The fieldwork introduced below corresponds to recent phases of data expansionism and operational consolidation. However, the implementation of RTPI in Dublin, particularly on bus services, evidences the long infrastructural timelines of ITS. Therefore, the section thereafter covers this initial period from the 1970s until the near-present, before then considering how RTPI standards were negotiated and how data-driven organisational change ensued.

Fieldwork and methods

Dublin's RTPI system covers Bus Átha Cliath (henceforth Dublin Bus), Bus Éireann (a public national coach company), Luas (tram system), and Iarnród Éireann (the public national railway company). Dublin Bus registered 136.3 million journeys in 2017, as compared to 37.6m for the Luas, and 45.5m for Iarnród Éireann nationwide (NTA, 2018: 201). The focus of research reflects the operational area of Dublin Bus, corresponding to the Greater Dublin Area (GDA). The GDA comprises four predominantly urban local authorities with a combined population of 1,347,359 in the 2016 census and three surrounding rural counties. There is no corresponding transport authority for the GDA. Instead, the national scale tends to be the favoured tier for integrated services across local authorities (Coletta et al., 2018a: 4). The National Transport Authority (NTA) is responsible for both national and regional transport planning including the GDA.

The choice of the fieldwork site and the topic of transport technology reflect a wider objective of tracking Dublin's self-promotion as a ‘smart city’ as part of a large multi-researcher project. It forms one of several complementary case studies on how smart technologies and Big Data are transforming urban life (cf. Cardullo and Kitchin, 2018; Coletta et al., 2018a; Perng and Kitchin, 2016). The empirical research presented in this article draws on 12 semi-structured interviews derived from two related fieldwork datasets. The first six are a subset are drawn from a larger set of 77 interviews conducted with government and city workers, corporations, and other stakeholders on Dublin's emerging smart city strategy in 2015/2016. These include five transport engineers from local and national government and one transport consultant, all of whom discuss ITS in relation to traffic control, road design and maintenance, and public transport reform. The remaining six are part of subsequent in-depth studies with transport operators (Bus Éireann, Dublin Bus, Luas) and traffic control room engineers (Dublin City Council) on interrelated RTPI and traffic management technologies in 2016/2017.

Interviews were conducted in situ in back offices (NTA, Luas, Bus Éireann) and control rooms (Dublin Bus and the traffic control rooms of Dublin City Council and South Dublin County Council). Observational fieldnotes from visits to the control rooms and impromptu conservations with operators provided additional perspectives. Interviewees were asked to explain their roles, the deployment of RTPI in their organisation and its relationship to other technologies, decisions on standards and their implementation, and their coordination with other transport providers and regulators. The insights offered by participants were supplemented by reports and secondary literature to inform the longer timeline of ITS in Dublin.

Four decades of pilots in Dublin

In Dublin as elsewhere, the development of ITS can be seen to be dependent on a supporting infrastructure which allows innovation to occur over multi-decadal timescales. There were many instances of transport innovation over the last few decades, but without sufficient coordination between the relevant bodies, they failed to scale up or attract sufficient continuity support from national or local government.

RTPI allows public transport users to consult the predicted departure time of services and determine the most efficient mode of travel, based on the latest information on traffic delays, interruptions and capacity. RTPI improves perceived reliability as passengers rate their transport providers more highly if they are kept informed of how the system is performing and can make more sophisticated information-driven travel decisions (Caulfield and O'Mahony, 2009; Watkins et al., 2011). Although transport operators can fall back on scheduled timetables and rosters to provide basic transport infrastructure, the higher levels of service attained with transport technologies are critically dependent on software and automation.¹

RTPI is dependent on automatic vehicle monitoring (AVM), a technology which reports real-time information on the location of vehicles back to a central server. AVM dates back to the late 1960s, with trials in the United States of America and other nations to develop radio-based triangulation systems to an acceptable degree of accuracy and temporal resolution. It was tested for select bus routes in urban areas throughout the 1970s with varying degrees of success (Roth, 1977). For mass transport systems, AVM promised a more efficient and less costly alternative to the ‘point men’, employed to record stop times of buses and inform schedule adherence and redesign (Roth, 1977). These elemental technologies of control contribute towards tackling the perennial issues of buses running ahead of schedule (understood to be considerably worse than running behind, as it causes disruption to drivers further back down the line and frustration to patrons) and maintaining even headway (where buses are distributed evenly along the route). It was recognised in the late 1970s that AVM could be used to provide both frequent and reliable service information to passengers and also integrate with traffic control systems to give signal priority to oncoming public transport vehicles running behind schedule (Symes, 1980: 237). Dublin Bus first trialled AVM in the 1970s using odometers fitted on buses that reported by radio to a central server every 45 seconds, subsequently rolled out to all bus depots by 1981. In 1985–1987, Dublin Bus also trialled traffic prioritisation based on a combination of infrared transponders installed on buses and roadside detectors. This was linked to the AVM system, but funding was not available to continue a successful pilot. The AVM system continued until the end of its useful life in the 1990s but was not replaced, with controllers reverting to radio contact with drivers (World Bank, 2011).

In 2001, a second trial of next-generation GPS-based Automatic Vehicle Location (AVL) called ‘Q-time’ was conducted on select Dublin Bus routes, and for the first time, RTPI signage was fitted (Caulfield and O’Mahony, 2004). This trial continued for three years during a period in which there was perennial threats (or opportunities) to liberalise the transport sector. The Irish Minister for Transport announced in 2002 that 25% of bus routes in the capital were to be opened to private competition, while also proposing the development of a ‘Dublin Land Use and Transport Authority’ in line with best practice across the European Union (Caulfield and O’Mahony, 2003: 2). Privatisation in transport is associated with market deficiencies including the dumping of non-profitable but socially important routes, and fare-creep on monopolised well-transited routes. Observing the UK experience, further inefficiencies may include competing companies serving the same routes, multiple and incompatible ticketing options, and the duplication of management and control resources (cf. Sørensen and Gudmundsson, 2010). Privatisation is therefore often accompanied with a parallel investment in new regulatory structures to mitigate these issues while proceeding on the ideological basis of lowering public investment costs and mitigating militant trade unionism (Gomez-Ibanez and Meyer, 2011).

In this context of uncertainty and privatisation in the early 2000s, Dublin Bus did not attain funding for a city-wide implementation of their Q-time pilot. Therefore, by the time RTPI was finally funded and implemented city-wide on buses from 2009 to 2011, it was a mature and largely consolidated technology commonplace in European cities like Gothenburg and Helsinki since the mid to late 1990s. Acting as a reverse salient, the arrival of GPS was making locational technologies part of everyday experience for both transport operators and personal devices (Kitchin, 2014b: 58), and could overcome technology resistances experienced during the first 1980s generation of AVM. The first service-wide implementation of RTPI in Dublin was for a new tram service in 2004, Luas, run on a franchise basis by Veolia, and before the bus implementation was finalised. It shares road-space with private vehicles for which it gets priority when approaching junctions. Sensors are placed at intervals of 100m or more which gather information from transponders on trams and send it to a central server and operations room.

As evident in the timeline below (Figure 1), Dublin Bus had accumulated or retained experience with RTPI through successive trials by the time of its definitive roll-out of AVL, in partnership with a specialist German transport technology company, INIT. A Dublin Transportation Office had been created in 1995 to put into place a transport strategy for the city region under the remit of the Department of Transport. It established a committee to oversee the implementation of RTPI in 2001, and contracted Atkins consultants in 2002 to create a general strategy. The report, published in 2006, noted the inconsistencies of information provision between services and the absence of an overarching mechanism to ensure holistic planning of both physical and informational infrastructure. It strongly recommended the creation of a specific public transport information office ‘with responsibility for collecting data, publishing information and setting standards’, ‘the development and marketing of a public transport “brand” common to all modes and operators that the public can identify with, trust and rely upon’, and ‘the development of a set of agreements and processes governing agency participation’ (Atkins, 2006: iv). The mooted idea of a land use and transport authority materialised in the form of the NTA, which fulfils the remit suggested in the report at the national scale. It also incorporated the Dublin Transportation Office and its data-modelling team (interview DSC27, NTA).

Figure 1.

A timeline of RTPI-related transport technology deployments in Dublin for bus and tram services, with further events on the left. Iarnród Éireann is not included, which relies on its legacy signalling systems in addition to AVL.

Figure 2.

Dublin bus live monitoring of schedule adherence in their control room.

As the NTA was not created until 2008, ‘it was agreed at the time [that the implementation of RTPI was initiated] that Dublin City Council was a better vehicle to actually procure the RTPI and subsequently that project and contract moved over to the NTA’ (interview DSC27, NTA). The NTA have adopted the infrastructure created by Dublin City Council and added further measures to ensure its resilience, while also extending RTPI and the Leap smart travel card to other cities and towns in Ireland.

The AVL data from services are fed back into a central RTPI server, and subsequently into roadside display panels via cellular communication or GSM, with data hosted in two data centres in the Dublin Docklands with various failsafe mechanisms. This informs two official NTA apps (RealTime Ireland and Journey Planner) that cover all services, as well as several operator apps (Iarnród Éireann and Luas). RTPI data are available via a public API to researchers and commercial developers under a CCBY 4.0 license. Programmers can write their own queries in XML (extended mark-up language) or JSON (JavaScript object notation) to pull down specific information, which can then be pushed into custom-made displays for specific purposes.

These developments have given rise to two forms of data expansionism: that within and between the core operators themselves and largely based on AVL, including scheduling, data analytics and traffic prioritisation; and that permitted by the API, for research and commercial usage. The next section details how RTPI standards are managed between operators and the regulator in order to support this two-tier ecosystem, while the final empirical section considers how AVL and RTPI are supporting the operational consolidation of data-driven functionalities.

Negotiating and maintaining data standards

It was the responsibility of Dublin City Council, and then the NTA, to create consistency of information provision, ensure infrastructure compatibility, and develop a common brand with its accompanying clear aesthetic. This involved the redesign of bus-stops, the rollout of RTPI displays for multiple operators, the creation of an efficient and reliable back-end and telecommunications system, and the policing of information systems to ensure interoperability between transport providers. This required a co-constitutive development of procedures encompassing both human operators and software, where the latter is understood as partial or fully automated procedural systems.

Anomalies and breakdowns make infrastructure visible (Star and Ruhleder, 2001), and their resolution allows us to follow how standards are negotiated and stabilised. Such is the case with ‘ghost buses’, where services shown in RTPI fail to materialise. This anomaly provides insight into the various orders of issues encountered and their relationship to organisational change as the NTA exercised its authority over transport operators. There are many reasons for ghost buses, one of which was due to communication failures between operators as discussed below and resolved by early 2016. Its elimination and that of other issues that affected accuracy involved both repairing bugs and policing adherence to standards. For 2017, it was reported that 97.5% of RTPI information was correct with reference to arriving within 1 minute of the ‘Due’ prediction (Bus Átha Cliath, 2017), up from 92% in 2012 (Worrall, 2012) and 89% when initially launched.

In agreement with language policies for State bodies, RTPI information, including destination information, on trains, buses, and trams, is displayed and announced in both English and Irish. The original versions of many Irish place names now share equal space with their anglicised counterparts, e.g. Cluain Saileach (a meadow of willows) and ‘Clonsilla’. The ghost bus anomaly resulted from a series of actions starting with Dublin Bus curtailing a bus and redirecting it before meeting its scheduled destination. Of the two information standards used by transport organisations in Ireland, VDV452 is used for scheduled timetables and SIRI for real-time feeds.² Dublin Bus changed a 19-character field in SIRI called ‘DestinationRef’ in a way that partners did not expect, as the employee cited below explains, because it did not have enough space for both languages, necessitating the use of a further field called ‘JourneyNote’:

For example, we send the journey ref [which] would say, ‘Okay, this journey is doing this destination and now that destination has changed’. So say, for example, a trip is going from A to B and we decide, okay, it is not going to B anymore, we'll bring it back somewhere. Now it can't handle that at the moment, the new destination, it still tells passengers you are going all the way to B. They are working on it and they are almost there. We had the same problem with the street signs for quite some time but they worked on that and they were fixed [by their external contractor] and that got resolved, that got changed. So now for example if you are going to Maynooth on a 66 [bus] for example, and for whatever reason, operational or whatever, they decide, okay this bus can't go the full journey, it can only go as far as Leixlip. We can now tell on the street signs, we can tell on our app on our website that to the passengers straight away, it could be on the bus itself, on the displays, they can be told this bus is now finishing in Leixlip. Whereas on the NTA's app they are still telling you ‘Maynooth’. So it will be fixed quite soon, they have been working on it for a number of months now so it should be fixed soon. But on the street signs and our web that is reflected correctly. But that is kind of a bug thing, that isn't anything to do with the standards really, that is more a bug on their internal system than anything. It is no limitation of SIRI or the VDV that has caused that. (Interview RTPI01, Dublin Bus)

Dublin Bus considered that their usage was consistent with SIRI and that the issue reflected an external contractor's inability to read this information. In contrast, the NTA interviewee cited below indicated ‘JourneyNote’ as being more for minor notes like being ‘guide dog-friendly’. This alleged non-standard use broke the information feed downstream, even if the change suited the challenge of frequently curtailing buses due to the large but necessary interruptions caused by Luas cross city.

A common solution for multiple language provision is to use a generic code in a look-up table with corresponding real-world names in various languages in adjacent columns:

Yes, and largely that works fine if you don't go putting stuff in places where it shouldn't be, like journey notes shouldn't be in there. So the system trying to understand that doesn't see it and therefore these, I suppose important things, because curtailments and cancellations are really what you need to know about in real-time. You need to know if your bus is not going to the destination or if it has been cancelled. So they are things we can fix but they are expensive because what you will find is that an incumbent will see that as an opportunity to, let's say, know that nobody is going to compete with their price and therefore I would say give in ridiculously high prices to do fairly simple stuff. […] And so really our experience is to kick back and say, ‘Sorry, guys’, right at the very start, stick to the specifications. […] So over the years the system has been in place and more people want to build ancillary systems or reuse the information, the more we have learned that right at the start you need to be the policeman. (Interview SD13, NTA)

The central regulator talks of their role of policing common standards to ensure fair competition between providers through shared rather than privileged information. It has weekly meetings with RTPI representatives from the main transport providers during which anomalies and their potential solutions are discussed. While assertive about its policing role, a solution was negotiated that allowed Dublin Bus a degree of flexibility in their implementation of SIRI, creating custom code to recognise Dublin Bus curtailments. The changes were visible first on the street signs, and reflected later on the NTA smartphone app.³

This second-order issue on adherence to common standards reflects a transition from putting in place the basic infrastructure towards consolidating RTPI in a new suite of procedures and processes, yet also interacts with higher third-order reconfigurations of power relations as the regulator imposes a new performance-driven ethos on operators. As the transport franchise-holder, the NTA has the power to outsource routes to other suppliers, oversees performance management for all contracted transport providers, and is therefore at the heart of negotiations on privatisation and standardisation. During the time of fieldwork, there were industrial disputes involving Luas tram drivers, and protracted discussions on the costs and benefits of partial privatisation of national services targeting Dublin Bus and Bus Éireann. Over the course of 2016 to 2018, 24 routes from Dublin Bus and 6 routes from Bus Éireann were listed for privatisation (Tallaght News, 2017) with UK firm Go-Ahead winning both competitions. The NTA manages routes, schedules, fares, vehicles, quality control and RTPI for new providers. In this manner, it hopes that like with Transport for London, tenders for routes are largely inconsequential for passengers through a model of competition for the market rather than in the market (Preston and Almutairi, 2013).

Operational consolidation here was the prerogative of an increasingly assertive state actor, the NTA, which addressed administrative fractures between State companies through control over both markets and data. This facet of data ratcheting comprises data-driven organisational control over operators and requires collective adherence to data policies. At the intra-organisational scale, within Dublin Bus and Bus Éireann, real-time data on the location of vehicles and drivers also enabled data-driven organisational control, in part based on new data-driven functionalities. This ratcheting provides momentum to the cycle of expansionism, experimentalism and consolidation as new datasets are created and further use-cases found.

Data-driven procedural change and data ratcheting

AVM and RTPI technologies in Dublin have given rise to a cascade of data-driven organisational changes. In many instances, these are local learnings of established international best-practices, yet also include instances of innovation such as the many small data-driven changes to routes to reach industry-standard RTPI accuracy. Dublin Bus's schedule management software is reliant on AVL and creates a more constant and intimate connection between drivers and controllers. The latter are housed in a control centre and seated at cubicles equipped with several graphical user interfaces for monitoring buses in real-time.

Much the same as the recording devices affixed to chain retail workers, these technologies constrain employee behaviour in order to provide an experience to customers that is invariable and therefore negotiable with less cognitive effort. Dublin Bus is a public company and many of the staff in the control room are ex-drivers themselves, their exchanges with drivers replete with banter and laughter yet nevertheless effecting a culture change within the organisation:

[B]efore the AVL system, the only way of knowing where a bus was, unless there was a guy out on the street watching the buses coming in and the inspector on the road. Or else it was the control up above calling the driver and saying, ‘Where are you?’ And now it is a case of … It was funny at first seeing it happen because they'd be saying I am in such and such a street. And the controller would say, ‘No you are not. I can see within 20 seconds of where you are’. That took a bit of getting used to and now the drivers are at the point where they are embracing it very much so. It took a while for them to trust it. […] But yeah, they have taken it on board now and they do realise all the old ways of operating have to change. (Interview RTPI01, Dublin Bus)

The interviewee above notes the role of AVL in supporting measures to improve quality of service. They are enforced not only by control room operators, on whose terminal screens buses are highlighted in red for being behind or ahead of schedule (see Figure 2), but also on vehicles themselves through devices to notify drivers if they are running ahead of their scheduled stops via a beep and a red dashboard light (only when the vehicle is stationary). Together with the control room and its real-time displays, data analytics, and a management team with key performance indicators to meet, these technologies all coerce human behaviour towards minimising error and irregularities. A realisation of the potential of data to drive internal reform has led to further local innovations, including using data analytics packages to recalibrate schedules (timetables) based on statistically examined journey times, and developing new internal performance metrics for management. These measures further improve RTPI accuracy, and evidence how data ratcheting reshapes internal organisational control as engineers experiment with, and operationalise, new functionalities.

In addition to schedule and route alterations, further interventions in the interest of efficiency can be either physical, such as changes to the road layout, or digital, through alterations to the junction signalisation patterns of traffic lights. The greater part of Dublin's road network is managed with industry-standard software (SCATS) that alters junctions dynamically in response to traffic by swapping between pre-programmed plans. Traffic prioritisation has now been rolled across the SCATS network for public transport by using RTPI data in SIRI format to estimate proximity (O'Donnell et al., 2018). This belated but eventually successful deployment of traffic prioritisation (see Figure 1) relates to third-order issues surrounding the dominant trend of transitioning to a low-carbon future by progressively revoking privileges once handed out willingly to the private motorist and increasing investment into public transport.

The schedule and fleet management software, the data analytics functions for route optimisation, and automated bus prioritisation form part of an established practice of ratcheting the recognised power of data to redefine procedures, and by extension, urban flows and spatial relations. Data-driven functionalities become perceived as integral during this phase of efficiency and rationalisation. A Dublin Bus technician notes that data are ‘growing more and more into decision-making around the company’ and adds:

So it is a complete [change of] mind-set, everything has changed. Once you trust in the data better decisions can be made […].

It was unbelievable straight away, things like when a door opened, you could see straight away, just information that you just couldn't have before so straight away you were just giving the power to the managers to see exactly what was going on, the routes, and ultimately that is getting back to the customers. (Interview RTPI04, Dublin Bus)

This localised data experimentalism and consolidation occur alongside larger-scale operational changes as centralised authorities coordinate transport services and data, creating new products and services to consolidate those initially created by individual transport providers. This includes the present BusConnects programme, initiated in 2017 and led by the NTA, for the rationalisation of all bus services in Dublin. Two NTA participants below recognise the potential of data for better strategic planning:

So I think once you have the kit with proper management and working between different companies you can make these improvements to raise the reliability and the perception of reliability to people. So I see expanding the system but also harvesting it to offer better products. (Interview SD13, NTA)

Yes, we use the data for a number of purposes. The initial objective was to provide more information to the consumers to actually make transport more accessible; that is the real value we saw from the information that we were collecting and providing. But certainly we have realised now that we have a huge wealth of information at our disposal in terms of where buses are at, the best journey patterns being used, and we can analyse that now and start using that to make the planning decisions more effective going forward. (Interview DSC27, NTA)

The expansion of metrics and accountability reinforces an ‘audit culture’ (Strathern, 2000), dependent on technologies that ostensibly serve passenger service provision but which also expand into technologies of control and regulation. Data ratcheting represents a quasi-Lamarckian evolutionary progression as organisations create ever more sophisticated secondary products from high-quality data, recognising its power and curating it with discretion according to socio-cultural specificities and legal frameworks. On the one hand, this may involve preventing such data from being misused for criminal purposes, while on the other, it could be used to maintain competitive advantage or increase control over consumers or citizens. The empirical data presented here account for largely internal and inter-organisational data-driven innovations yet may eventually lead to broader data assemblages beyond the transport sector, aligning with broader commercial or political interests.

Conclusion

In their future roadmap of ubiquitous computing, Dourish and Bell (2011) highlight the emergent nature of technologies, which initially developed separately, coalesce in what are seen retrospectively as unitary systems such as the smart home or the city dashboard (Kitchin et al., 2016). Technological changes are also essentially organisational and procedural, liberating the human mind from repetitive drudgery, and facilitating new forms of behaviour between people and their interaction with the environment. Tracing their evolutionary development reveals the multiple path dependencies and reverse salients (Hughes, 1993) that characterise their real-world implementation, and furthermore illustrates the value of blending the historicism of infrastructure studies with the study of data assemblages.

The rollout of RTPI in Dublin is interrelated with broader transport technologies including automated vehicle location systems and traffic control systems, and evidences how its local implementation is part of an arc of technological changes in ITS. It initiates in the 1970s with advanced trials that were inconsistently resourced by the State until the context itself had shifted. The national economy had moved towards higher-value technologies and services with a more informed and demanding workforce, necessitating the provision of seamless multimodal transport, integrated ticketing, and RTPI.

The deployment of RTPI allowed individual operators such as Dublin Bus to experiment with and drive their own internal procedural reforms, with instances of both localised learnings and more genuine innovations as local actors participated in European and international networks for developing and disseminating best practices. These activities are described here as instances of data ratcheting, as new data-driven functionalities are implemented iteratively as they are discovered in a local context, oriented towards a more rationalised data-driven culture. Glitches such as ‘ghost buses’ provide a window into the politics of data, as the transport sector modernised and the NTA consolidated its mandate to improve passenger experience by policing standards and controlling providers. This hierarchy of procedural authority, from regulator to operator to driver, has been strengthened through the creation of governmental bodies that ensure seamless exchange of data and the operational consolidation of data-driven procedures. In addition to favouring a rationalised audit culture (Strathern, 2000), this will likely assist in future data-driven functionalities that integrate with other sectors and services.

Urban datafication in transport has been characterised here as consisting of the three translations of data expansionism, data experimentalism and operational consolidation. It is tentatively argued, pending further comparative reviews of in-depth local studies, that there will be an increasing tendency to curate and streamline data as the potentialities of Big Data shared across multiple domains are appreciated and realised. Transport operators, particularly those serving massive urban populations, can use tracking technologies for efficiency, commerce and security. Transport for London, for example, while maintaining their open data portal and subject to comprehensive UK and EU data protection and privacy regulations, are investing in large-scale tracking technologies to inform their services (McMullan, 2018; Sweeney, 2018). Transport data may be combined with social networks and public services profiles to perform new forms of citizenship as smartphone-delivered personalised notifications and services become normalised. Further research could enquire as to the multiple data streams that are becoming operational, their epistemologies, and their primary benefactors. For instance in China, transport apps for citizens are being tied to unique identifiers that may restrict or incentivise transport options according to their government-measured ‘social credits’ (Carney, 2018). Such developments may similarly benefit from the comprehensive approach of data assemblages that contextualise data-driven technologies in their socio-cultural contexts, as well as from the extended timescales of infrastructure development and attention to glitches and breakdowns.

Footnotes

Acknowledgements

I am indebted to Professor Rob Kitchin for his very helpful feedback and input into this article and research,which formed part of a larger research project on the smart city for which he was principle investigator. My thanks also to Dr Sheila Convery for her comments on the article,and to the editorial team and three anonymous reviewers for their comments.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: The research for this article was provided by a European Research Council Advanced Investigator Award,‘The Programmable City’ (ERC-2012-AdG-323636).

Notes

References

Atkins (2006) A Strategy for Integrated Public Transport Information, including Real Time, in the Greater Dublin Area. Dublin, Ireland. Available at: www.nationaltransport.ie/downloads/archive/atkins_report_2006.pdf (accessed 7 August 2018).

Bus Átha Cliath (2017) Bus Átha Cliath Annual Report and Financial Statements 2017. Available at: www.dublinbus.ie/About-Us/Reports/Annual-Reports/ (accessed 11 April 2019).

Cardullo

Kitchin

(2018) Being a ‘citizen’ in the smart city: Up and down the scaffold of smart citizen participation in Dublin, Ireland. GeoJournal 84: 1–13.

Carney M (2018) Leave no dark corner. ABC News, 18 September. Available at: www.abc.net.au/news/2018-09-18/china-social-credit-a-model-citizen-in-a-digital-dictatorship/10200278 (accessed 21 September 2018).

Caulfield B and O'Mahony M (2003) Real time passenger information: The benefits and costs. In: The association for European Transport Conference, Strasbourg, 8 October 2003. Available at: http://www.tara.tcd.ie/bitstream/handle/2262/79614/Caulfield O'Mahony%2C Real time passenger information costs and benefits.pdf (accessed 29 July 2019).

Caulfield B and O'Mahony M (2004) The provision of on street passenger information via real time passenger information: A case study of Dublin. In: 12th IEE international conference on road transport information and control, 2004. RTIC 2004. April 2004. DOI: 10.1049/cp:20040001.

Caulfield

O'Mahony

(2009) A stated preference analysis of real-time public transit stop information. Journal of Public Transportation 12(3): 1–20.

Coletta

Heaphy

Kitchin

(2018a) From the accidental to articulated smart city: The creation and work of ‘Smart Dublin’. European Urban and Regional Studies 1–16. Available at: https://doi.org/10.1177/0969776418785214 (accessed 29 July 2019).

Coletta

Heaphy

Perng

S-Y

, et al.(2018b) Data-driven cities? Digital urbanism and its proxies: Introduction. TECNOSCIENZA: Italian Journal of Science & Technology Studies 8(2): 5–18.

10.

Dourish

Bell

(2011) Divining a Digital Future: Mess and Mythology in Ubiquitous Computing, Cambridge, MA: MIT Press.

11.

Edwards

(2010) A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming, Cambridge, MA: MIT Press.

12.

Foth

Brynskov

Ojala

(2015) Citizen's Right to the Digital City: Urban Interfaces, Activism, and Placemaking, Singapore: Springer.

13.

Giffinger

Pichler-Milanović

(2007) Smart Cities: Ranking of European Medium-sized Cities. October, Vienna: Vienna University of Technology.

14.

Gomez-Ibanez

Meyer

(2011) Going Private: The International Experience with Transport Privatization, Washington, DC: Brookings Institution Press.

15.

Gray

Gerlitz

Bounegru

(2018) Data infrastructure literacy. Big Data & Society 5(2): 1–13.

16.

Hughes

(1993) Networks of Power: Electrification in Western Society, 1880–1930, Baltimore, MD: The Johns Hopkins University Press.

17.

Kitchin

(2014a) Big Data, new epistemologies and paradigm shifts. Big Data & Society 1(1): 1–12.

18.

Kitchin

(2014b) The Data Revolution: Big Data, Open Data, Data Infrastructures & Their Consequences, London: Sage.

19.

Kitchin

Coletta

Heaphy

(2018) Actually existing smart Dublin: Exploring smart city development in history and context. In: Karvonen

Cugurullo

Caprotti

(eds) Inside Smart Cities: Place, Politics and Urban Innovation, London: Routledge, pp. 83–101.

20.

Kitchin

Dodge

(2011) Code/Space: Software and Everyday Life, Cambridge, MA: MIT Press.

21.

Kitchin R and Lauriault TP (2014) Towards Critical Data Studies: Charting and Unpacking Data Assemblages and Their Work. ID 2474112, SSRN Scholarly Paper, 30 July. Rochester, NY: Social Science Research Network. Available at: http://papers.ssrn.com/abstract=2474112 (accessed 12 December 2018).

22.

Kitchin

Maalsen

McArdle

(2016) The praxis and politics of building urban dashboards. Geoforum 77: 93–101.

23.

McMullan T (2018) Transport for London to start automatically tracking bus passenger numbers. Alphr, 7 June. Available at: www.alphr.com/go/1009545 (accessed 12 November 2018).

24.

Meinel

(2004) Molecules and croquet balls. In: Chadarevian

Hopwood

(eds) Models: The Third Dimension of Science, Illustrated edition. Stanford, CA: Stanford University Press, pp. 242–275.

25.

NTA (2018) Annual Report & Financial Statements 2017. Rialtas na hÉireann. Available at: www.nationaltransport.ie/wp-content/uploads/2018/12/NTA_AR_-_FS_2017_ENGLISH_FINAL.pdf (accessed 11 April 2019).

26.

O'Donnell M, Bolger N and Calvey J (2018) Data driven decisions for centralised bus priority in Dublin City. In: JCT Traffic Signals Symposium & Exhibition, Nottingham, September 2018, p. 15. Available at: www.jctconsultancy.co.uk/Symposium/Symposium2018/symposium2018.php (accessed 10 December 2018).

27.

Perng

S-Y

Kitchin

(2016) Solutions and frictions in civic hacking: Collaboratively designing and building wait time predictions for an immigration office. Social & Cultural Geography 19: 1–20.

28.

Preston

Almutairi

(2013) Evaluating the long term impacts of transport policy: An initial assessment of bus deregulation. Research in Transportation Economics 39(1): 208–214.

29.

Reed

(2018) Discovering the data-driven city. Breakdown and literacy in the installation of the Elm Sensor Network. TECNOSCIENZA: Italian Journal of Science & Technology Studies 8(2): 81–104.

30.

Reed

Webster

(2010) Architectures of motility: ICT systems, transport and planning for complex urban spaces. In: Nunes Silva

(ed) Handbook of Research on E-Planning: ICTs for Urban Development and Monitoring, Hershey, PA: Information Science Reference, pp. 365–387.

31.

Roth

(1977) History of automatic vehicle monitoring (AVM). IEEE Transactions on Vehicular Technology 26(1): 2–6.

32.

Shapiro

(2018) The urban stack. A topology for urban data infrastructures. TECNOSCIENZA: Italian Journal of Science & Technology Studies 8(2): 61–80.

33.

Shelton

(2017) The urban geographical imagination in the age of Big Data. Big Data & Society 4(1): 1–14.

34.

Shelton

Zook

Wiig

(2015) The ‘actually existing smart city’. Cambridge Journal of Regions, Economy and Society 8(1): 13–25.

35.

Sørensen

Gudmundsson

(2010) The impact of governance modes on sustainable transport – The case of bus transport in Greater Manchester, UK. World Review of Intermodal Transportation Research 3(1): 8–25.

36.

Star

(1999) The ethnography of Infrastructure. American Behavioral Scientist 43(3): 377–391.

37.

Star

Bowker

(2006) How to infrastructure. In: Lievrouw

Livingstone

(eds) Handbook of New Media: Social Shaping and Social Consequences of ICTs. Updated student edition, Thousand Oaks, CA: Sage Publications, pp. 230–245.

38.

Star

Ruhleder

(2001) Steps toward an ecology of infrastructure: Design and access for large information spaces. In: Yates

van Maanen

(eds) Information Technology and Organizational Transformation: History, Rhetoric, and Practice, Thousand Oaks, CA: Sage Publications, pp. 305–346.

39.

Strathern

(2000) Audit Cultures: Anthropological Studies in Accountability, Ethics, and the Academy, London: Routledge.

40.

Sweeney R (2018) Opening data fully to improve London's transport network. Intelligent Transport, 26 February. Available at: www.intelligenttransport.com/transport-articles/65404/opening-data-full-capacity-transport/ (accessed 12 November 2018).

41.

Symes

(1980) Automatic vehicle monitoring: A tool for vehicle fleet operations. IEEE Transactions on Vehicular Technology 29(2): 235–237.

42.

Tallaght News (2017) Sean Crowe has described the privatisation of 24 Dublin Bus routes as worrying, 18 August. Available at: www.newsgroup.ie/sean-crowe-has-described-the-privatisation-of-24-dublin-bus-routes-as-worrying/ (accessed 9 April 2018).

43.

Watkins

Ferris

Borning

, et al.(2011) Where is my bus? Impact of mobile real-time information on the perceived and actual wait time of transit riders. Transportation Research Part A: Policy and Practice 45(8): 839–848.

44.

Wiig

(2019) Incentivized urbanization in Philadelphia: The local politics of globalized zones. Journal of Urban Technology 1–19. Epub ahead of print 2019. Available at: https://doi.org/10.1080/10630732.2019.1573628 (accessed 29 July 2019).

45.

Williams

(2008) Intelligent Transport Systems Standards, Boston, MA: Artech House.

46.

World Bank (2011) Intelligent Transport Systems – Dublin, Ireland. Available at: www.ssatp.org/sites/ssatp/files/publications/Toolkits/ITS%20Toolkit%20content/case-studies/dublin-ireland.html (accessed 13 December 2018).

47.

Worrall JJ (2012) For travel innovators, timing is everything. The Irish Times, 5 July. Available at: www.irishtimes.com/business/technology/for-travel-innovators-timing-is-everything-1.529501 (accessed 13 December 2018).

48.

Yates

(1993) Control through Communication: The Rise of System in American Management, Johns Hopkins paperbacks ed. Studies in industry and society 6. Baltimore, MD: Johns Hopkins University Press.