Smart Cities and M³: Rapid Research, Meaningful Metrics and Co-Design

Abstract

The research described in this paper is undertaken under the banner of the smart city, a concept that captures the way urban spaces are re-made by the incursion of new technology. Much of smart is centred on converting everyday activities into data, and using this data to generate knowledge mediated by technology. Ordinary citizens, those that may have their lives impacted by the technology, usually are not properly involved in the ‘smartification’ process. Their perceptions, concerns and expectations should inform the conception and development of smart technologies at the same extent. How to engage general public with smart cities research is the central challenge for the Making Metrics Meaningful (MMM) project. Applying a rapid participatory method, ‘Imagine’ over a five-month period (March – July) the research sought to gain insights from the general public into novel forms of information system innovation. This brief paper describes the nature of the accelerated research undertaken and explores some of the themes which emerged in the analysis. Generic themes, beyond the remit of an explicit transport focus, are developed and pointers towards further research directions are discussed. Participatory methods, including engaging with self-selected transport users actively through both picture creation and programmatically specific musical ‘signatures’ as well as group discussion, were found to be effective in eliciting users’ own concerns, needs and ideas for novel information systems.

Background to and Overview of the Research Context

The research described in this paper was undertaken by a team which was thrown together at a ‘Sandpit’ event at the Open University on the 8th February 2016. The team members were from different Faculties and many of them had not met before the 8th. In four hours the team formed, came up with an idea for research, presented that idea to gate-keepers and were awarded funding. By the 20th July the research was finished.

Between the 8th February and the 20th July the team firmed up their initial research idea – about how metrics and data more generally can be made more meaningful to the public – into a focused assessment of public engagement with public transport information provision in smart cities. The research has involved seven active academics and has engaged over 30 members of the public in two workshop events in Milton Keynes. The research has innovated a fresh interpretation of the visualisation diagram, the ‘Rich Picture’, and provided an opportunity for the public to translate their wishes related to a better transport in the city into ideas for an innovative technology co-designed by them.

In outline, in this paper we set the context for public transport in Milton Keynes, the nature of Smart Cities and the systemic issues which they evoke. Building on this the article explores the literature around transport and the implications of non-inclusion for citizens. In the context of the literature our research question emerges with clear focus on the citizens’ requirement for transport knowledge. We then describe our research process, the methods engaged and the outcomes of the Action Research Workshops engaged. Finally, we conclude with some modest recommendations.

Context – Public Transport in Milton Keynes

The research is located in Milton Keynes (MK), in the United Kingdom. The intentions of the study were both to build upon existing knowledge of smart cities (see below), but also to reflect on matters of concern specific to the city. MK is a small but growing city with a population in 2016 of around 230,000. The city is located near London and expecting both population and economic growth in the future. The city faces two issues pertinent to its public transport system: its physical infrastructure and the place of public transport in the political life of the city.

MK is a new city, famously, and has a very particular form of infrastructure. It is a car centred city which operates through an 11 × 10 grid system of arterial roads on which traffic tends to run at motorway speeds. Those grid roads are placed at approximately 1 km intervals and the spaces between them hold residential areas. These are formed of estates, which are traversed through networks of low speed, often curved, residential streets, accessed from only one or two points on the grid roads. Long distance pedestrian and cycling provision is available through the “Redways¹”, routes free of motorised traffic that broadly follow the grid roads. The city’s road infrastructure creates unique challenges to public transport in MK. Unlike cities which follow the classic Victorian design, MK’s main roads do not also operate as local centres. This makes the ‘hubbing’ effect that public transport typically relies upon less pronounced in MK. The distributed population combined with the longer distances which must be covered by buses to access residential areas via the grid roads present particular challenges.

Moreover, public transport use has tended, until recent years, to be regarded as a secondary priority to private car use, and public transport users have observed that the provision is often inefficient and unreliable. MK is known as a car dependent city (http://www.bettertransport.org.uk/media/10-december-2014-car-dependency%20research), though the 2011 census reveals that nearly 19% of the city’s households do not have access to a car (18,656 households of a total of 98,584 – see http://www.ons.gov.uk/ons/rel/census/2011-census/key-statistics-for-local-authorities-in-england-and-wales/rft-table-ks404ew.xls). Thus there is a need for a more focused and holistic approach to public transport to improve mobility.

Context – Systems and Smart Cities – But where are the People?

Transport systems often emerge with unintended consequences due to a lack of holistic forethought. In this journal Khirsty and Zeitler identified some of the challenges which mobility provide in a systemic context (Jotin Khisty and Zeitler 2001). Transport is clearly an issue/ concern which, by its inter-connectedness and multiple layers provides an opportunity for systemic problem structuring (Ülengin et al. 2010) and a participatory approach (Farkas et al. 2014; Carlsson-Kanyama et al. 2008).

One of the requirements for our research was to be systemically clear on terms and meanings in this confusing area. For example, there is a lot of academic work that goes on under the banner of ‘smart cities’; which is ironic given that much of that work begins with a short discussion about the fluidity of the term smart city itself.

“The term ‘smart city’ has been variously defined within the literature…” (Kitchin 2014: 1).

This quote from Kitchin is one example of many. Broadly, attempting to avoid a lot of detail, this is because the terminology of smart city preceded many of the technologies today defined as ‘smart’. Thus, as one version of ‘smart’ was created, the technology moved on and new technological practice was retrofitted into a pre-existing landscape of terminology. Likewise, smart has a huge range of proponents from sectors whose cultural, commercial and political priorities differ widely, each of whom lend their support to a version of smart which corresponds to their particular interests.

The research described in this paper is undertaken under the banner of smart city research too. For clarity, in this project we employ a definition of the smart city with a wide purview, understanding it as a piece of conceptual terminology which explores the way urban spaces are re-made by the incursion of new technology and/or the digital.

The production and usage of data is central to smart cities. Almost always quantitative, data are the tool by which the messiness of human / urban life can be made meaningful to a computer system. Much of smart is centred on converting everyday activities into data. And transport is one of those central places where this occurs.

Literature Reviews of the Transport Data Landscape Reveal a Focus on Two Areas

The first is the increasing use of smart ticketing technologies like the Oyster card in mega-cities such as London (for a wider debate see: Coskun et al. 2014). These technologies allow very detailed collection of data on transport users’ activities as they move around the system (Nishiuchi et al. 2013). In other situations, where expensive technologies like smart ticketing are not available, citizens are enrolled to provide data on their movements. Often incentivised in some way, citizens activate an app on their phone which tracks their movements through the phone’s location services and feeds back to the app’s creator (which could be either the transport network itself, or it could be an independent body). These apps have been used to measure, for instance, the real time movement of trams or buses so that public transport users can receive live modifications to a printed timetable. Similarly, Google Maps’ traffic app uses the real time movements of Google Maps app users to identify and label traffic blockages on the roads (for a compendious discussion of the area see Mac-Gillavry 2013).

Two things are of interest for the research described here.

• First, as is perhaps inevitable in the early use of a technology, the data landscape has focussed on what it is possible to collect rather than what is necessarily required by transport users. This means that there is a natural tendency for data to be moulded to a use rather than the use being identified first and data being collected to answer that use.

• Second, data tends to be funnelled upwards to public transport operators (such as Transport for London or app developers). These operators may choose to make data available, but they may not necessarily do so. This is not automatically a world of open data.

Tomitsch and Haeusler (2015) talk about infostructures as being complementary to infrastructures. By this, they mean that collection and deployment of good data can help to enhance use of urban infrastructure and do so more cheaply. In simple terms, a new train line is expensive to construct, while providing people with data that shows when the existing infrastructure is less busy and encourage people to plan their travel accordingly could be a substantially cheaper means to reduce the pressure on infrastructure and improve the quality of travellers’ journeys.

That said, we also know that people do not simply make better decisions because there is more data available. The factor which is most influential in whether urban dwellers make use of an expanded data landscape that makes public transport use easier and more efficient is whether they are accustomed to using public transport already (Farag and Lyons 2010, 2012). Providing new data and better ways for people to navigate the system is useful, but, if there is a policy incentive to encourage people to use public transport, better technology is not what is going to achieve that. People will need to be encouraged onto the transport network first of all. This raises the question of what role citizens play in the underlying data structures. Participation can be inclusive and there are degrees of inclusivity. For example, in June 2016, the Milton Keynes “Hackathon” supported by the ‘MK Smart’ research project ran a weekend event where users and app developers got together voluntarily over a weekend to work on an app which will help people navigate the Redways’ in Milton Keynes.

It can be argued that if research involves a detailed and comprehensive use of metrics then it will tend to be experienced as excluding of the general public – the metrics ensure that there will be a cognitive and cultural barrier to wider participation – largely due to the exclusivity of access to and familiarity with technical knowledge. We endeavoured to break down these barriers to exclusivity by adopting a co-design or participatory design research methodology. This is a mode of design practice in which the users are participants in design practice as well as the design experts (from business, academia, and so on).

Participatory inclusion or ‘co-design’ in research processes is an established, multi-disciplinary concept, having been practiced in a number of fields for a good few decades (see for example: Chambers 1995; Fraser et al. 2006; Buscher et al. 2002; Cinderby 2010), and it is also a design concept which is gaining traction as commerce becomes more rooted around experience than physical goods as products – Sanders and Strappers (2008) elaborated on the co-design process as shown in Fig. 1.

Fig. 1

Co-design. (Source: Sanders and Stappers 2008, p. 6)

The process works by (following the chart along) bringing people in at the beginning of the product cycle (before the designers even know what they want to make, more or less), allowing people to have ideas that might pull the design process in a range of directions before, later on, using both them and the design team to hone those ideas. Not restricted to desirable technical features, the ideas discussed usually also reveal constraints related to current regulation, market organisation, among other formal constraints, and more social and informal aspects such as personal values, habits, local culture, and so on.

This sociotechnical and more holistic approach underlay our research method:

Research Question, Approach and Focus

Given our concerns regarding the top-down and expert-driven nature of existing metrics and our interest in involving the public in co-design, our initial research began with a simple question for public transport users in Milton Keynes:

“What information issues do you encounter when you consider the transport options you have to make on a daily basis?”

Following a couple of iterations, we refined this question in such a manner as to be easy to grasp and less off-putting in terms of subliminal assumptions (e.g. regarding existing knowledge/ expertise, etc.). The final form of our question was:

“What do I Wish I’d Known before I set Out or during My Journey?”

Early on the research team adopted a research approach and focus in order to explore this question. The approach and focus is set out in Fig. 2. At this stage a specific research application had not yet been defined, the reference to agent-based modelling in Fig. 2 indicates a potential link to detailed predictive modelling analysis with a person-centric, ‘bottom-up’ modelling paradigm matching that of the Imagine methodology used in the co-design process described here.

Fig. 2

Research approach and three key areas of focus

The approach combines three elements in two dynamic combinations:

• The first element is metrics and analytics for transport in smart cities

• Secondly the mediation and translation of these metrics and analytics – assumed to involve an ‘app’ of some kind.

• Thirdly, the users of the translated metrics.

In the second part of the Figure these three are shown in reverse order. Thus metrics informs application and this is provided to users/ publics. But, at the same time the public provides demands on the nature of the app and this in turn determines the data requirement.

For both metrics and users/people, the research team were interested in exploring narratives and stories which set out the main concerns and needs of the public for information on journeys involving public transport (on the one hand) and the limitations and uses of data (on the other).

As urban communities become smarter, data plays an increasingly significant role in their organisation and management. But neither as an object nor in the processes of its collection should data be considered value free or objective (Kitchin 2014). Instead, data can be understood as both partial and fallible. Knowing this, the aim of this exploratory project was to work with members of the public in MK to identify how they were served, or not, by the existing data landscape in relation to transport, and particularly public transport, in the city. In doing so, our aim was to provide members of the public with the opportunity to take a lead in considering their transport requirements and what improvements they would want to see applied to that data landscape. Following Tomitsch and Haeusler (2015), it is helpful to understand cities as being composed of physical infrastructures which it is now possible to enmesh with data led ‘infostructures’. These infostructures make it possible to enhance the utilisation of existing physical infrastructure, a far cheaper and less resource intensive proposition than new material constructions. In a city like MK, where infrastructure is not currently stretched but the population is projected to grow by 20% between 2012 and 2026,² now is an appropriate moment to be considering such questions. This brief review considers two key areas discussed in the literature.

• First, it looks at the workings of the existing transport data landscape;

• Secondly, it examines the way transport users are able to participate in bringing about the tools to utilise or manipulate this data.

The Existing Data Landscape

Data plays a critical role in urban transport planning and has received considerable attention in the literature. It is known that transport data is more likely to be available in urban areas, and that newer technologies, such as real time updates, are also more commonly employed in cities (Velaga et al. 2012). A key question for this project surrounds the way by which that data is collected, and for whom it is collected. A considerable amount of data is accumulated from new smart payment systems for urban transport networks, and there are various pieces of work which assess the functionality, utility, and ethical implications of the accumulation of that data (Allwinkle and Cruickshank 2011; Avoine et al. 2014; Bagchi and White 2005; Pelletier et al. 2011). Echoing broader moves in smart city activity towards crowdsourcing or citizen sensing (Gabrys 2014; Salim 2012), where smart ticketing systems are not available, the users of public transport have been enrolled to generate data using their own devices (Farkas et al. 2014, 2015; Thiagarajan et al. 2010). Interestingly, these existing studies of data accumulation tend to be centred on what kind of data it is possible to accumulate rather than what kinds of data would be useful for public transport users. Instead of seeking to find data to fit the problems, what our review appears to suggest is that data is generated and the learning it makes possible is retro-fitted to the problems. While technological development traditionally plays an important role in leading the creation of new, unimagined possibilities and markets, our review nevertheless implies that there is an important gap in the literature which is centred around user-driven demand for data. This is a gap which the current research has sought to address. The question for research (such as that contained in the M³ project) begins, not with what data it is technically possible to collect, but with an assessment of what users of transport systems require. Some examples of the utility of this approach exist. For instance, in Dhaka, crowdsourced data was employed to produce the first public transport map in a city in where, though many citizens would find a map useful, a lack of formal organisation of the public transport network had made such a map unrealisable (Zegras et al. 2015).

The question of how to produce data which directly responds to transport users’ requirements is significant given what we know about how data influences the way travellers make use of transport networks. It is known that people may not make personal transport choices that fit with their desires for the urban areas they inhabit – taking the car while simultaneously wishing for a less polluted city, for instance– and it is thought that the provision of better transport data might reduce such instances by facilitating increased public transport use (Lisson and Hall 2016). However, other research has found that while better data provision enhances the experience of public transport use amongst existing public transport users, the provision of data on its own does not encourage users on to public transport who are not already habituated with the networks (Farag and Lyons 2010, 2012). The ways in which public transport data used by travellers is also important. Public transport data is employed by the public to reduce travel time and affective effort or stress while en route, whereas information which will help reduce the cognitive or physical effort of travelling is collected before departure (Grotenhuis et al. 2007). Data may also be used by travellers to alter their route choices as they are travelling (Trozzi et al. 2015).

Transport Users Ability to Participate in Bringing about the Tools to Utilise or Manipulate this Data

Experimental work on how to deliver the data accumulated is present in the literature. In this review, we have briefly surveyed a series of articles which describe how apps are created, though these focus on the technical aspects of app development undertaken by professionals (Barbeau et al. 2014; Mirri et al. 2014; Nuzzolo et al. 2013). Also illustrative is work which has examined the best way to present QR codes³ on posters located in bus stops in order to direct travellers to a website which delivers information on forthcoming buses (Gammer et al. 2014). Citizen co-created applications (Desouza and Bhagwatwar 2012) exist in the world of transport use (Baraniuk 2013; Anon 2016; Paradi-Guilford et al. 2013), some of them designed with the purpose of persuading more sustainable transport choices (Mitchell et al. 2015). While this has opened up the arena from a purely top down approach, what is not addressed by this relative democratisation is the possibility of participation by those who do not have the time or technical skills to undertake app creation themselves. The M³ project has had this mode of engaging publics at its heart.

To conclude, then, this literature review has highlighted two key points within which the M³ study is located. The first is that transport data exists and is valuable, but that commonly the kinds of data which are collected are those which existing or new infrastructures make possible, rather than those specifically designed to be of relevance to travellers. We have come across no projects which have attempted to do what we have done; rather, approaches to data seem to be largely top down. Yet we know, from this research, that users have particular requirements from data which might not necessarily be best served by the kinds of data which are currently collected. Second, and building on the first point, this literature review has demonstrated that spaces for citizen participation are limited to those with the technical skills to undertake technical developments. While much data is made open, often via the General Transit Feed Specification also known as the.GTFS file type⁴ there is little scope for wider public participation, and examples of lay publics being directly enrolled in decision making processes in the way addressed by this research were not found.

M³ Project Process and Method

At the outset, the M³ project team identified a series of stages to the project process. Broadly speaking these were as follows:

• Initial workshop with local users:

  • identify and engage local users concerns regarding meaningful metrics for transport decision making in Milton Keynes.

  • Assess the data requirements for the specified concerns/ issues

  • Assess the technology / app issues for the specified concerns

• Preliminary analysis of user input, distillation of main themes, musical signature composition etc.

• Second workshop with local users:

  • identify potential means to realise technologies which can assist with transport decisions in terms of user interface design, data and socio-technical requirements.

Key to the process was a means to engage the public in Milton Keynes and for this the Imagine method was selected. Imagine (Bell and Morse 2010) is a community engagement method broadly based upon the Soft Systems Method of Checkland, but deliberately ordinated around the Kolb learning cycle and previously applied widely in environmental contexts – specifically relating to the application of metrics to issues of local sustainability (Checkland and Poulter 2006; Kolb 1984; Bell et al. 2013). The Imagine method was developed as a means to engage local populations in co-design of sustainability measurement and in problem structuring (Pidd 2011; Bell and Morse 2013) by engaging this population in the consideration of what is important, what is the current state of important things, and how they can be measured. The objective of the method is to gain information on data and indicators from mixed lay/ technical communities. It was always intended that Imagine be a methodology in the fullest sense of the term, evolving with time and capable of being cut and edited to meet requirements of local context.

Notwithstanding this flexibility, the key element of the method is the Rich Picture or RP (for a fuller description see Bell et al. 2016):

“The RP has been around for many years and is a cartoon-like diagram which encourages people to apply what Rose Armson brilliantly described as ‘optimal indiscretion’ to explore complex issues and, in many cases, set out realities and truths which, for one reason or another, cannot easily be spoken about or written formally. RPs contain this ‘release’ function, providing indiscreet insight into troublesome areas. We argue that the RP drawing enriches problem-solving and, in the long term, saves time and resources from being expended on erroneous and/or superficial tasks. RPs embody the commonly expressed view that ‘pictures paint a thousand words’.” (Bell et al. 2016, page x)

The RP is a device for gaining spontaneous and unguarded thoughts (as well as considered and guarded ideas) from groups of around 5–10 stakeholders. A RP is shown in Fig. 3.

Fig. 3

A Rich Picture drawn in Lebanon

A RP is a free form diagram with few rules other than to use as few words as possible. It is produced by a group on a large poster sheet of paper with coloured marker pens. This sample RP is taken from a project relating to the sustainability of the coastal fringe that took place in Lebanon. It is one of four produced at the same time by different stakeholder groups. Such a picture (Fig. 3) can be assessed by a form of Content Analysis, as the description in Table 1, below, illustrates.

Table 1. A content analysis of Fig. 3

To explain the content analysis is a little more detail:

• Column 1 contains the major themes drawn out by the research team assessing this and any other RPs produced at the same time in the same workshop. These themes are common to all RPs.

• Column 2 contains the specific relationship between this RP and the themes in column 1.

• Column 3 contains the themes mentioned by Stakeholders when they described their RP in a summarising session following the drawing process. The colour coding relates to similar themes which emerged across the other RPs which emerged in the workshop.

The penultimate row contains the plenary reflections summarising the RP as provided by the Stakeholders.

The final row is the research team’s overall assessment of all the previous information.

It contains the unique ‘signature’ of the RP which draws upon all the previous information. This is produced by the research team but can be developed in collaboration with the workshop participants.

This approach to Content Analysis is called Eductive Interpretation Content Analysis or EICA (for an overview of the approach see: Bell et al. 2015, 2016).

It is suggested that from the combination of interpretations of the original RP as contained in the EICA it is possible to educe the main themes of concern to the group who draw the RP. The EICA, when provided back to the participants in the workshop, can provide the basis for progress. In the case of M³ it was planned that a supplementary, second workshop should follow. At this workshop the original participants would be:

• Apprised of their original work

• Shown the EICA based signature for the first workshop as a whole

• Provided with a musical equivalent of the EICA based signature – this was an innovation on the method and was intended to provide a further creative catalyst to the participants

• Encouraged to design a technology which help to address the issues of concern noted in the first workshop.

Co-Designing Considering Socio-Technical Requirements

In line with the Socially-Aware Computing approach by Baranauskas (2014), this project understands technology design beyond defining technical requirements only.

Instead, design is considered a process originated in the society, embracing and surpassing informal aspects (e.g. peoples’ values, beliefs, habits, etc.), and some formal ones (regulation, metrics, for instance), towards the construction of a technical system. The technical system, on the other hand, impacts the formal and informal levels, towards influencing people in the society. Creating a new technology demands the articulation of meanings in the social group also in the formal and informal aspects. Figure 4, adapted from Baranauskas (2014), illustrates that.

Fig. 4

The Socially-Aware Design Approach (From: Baranauskas 2014)

The Socially Aware Approach considers the involvement of social group with a diversity of experiences and worldviews in participatory design practices, including community leaders and other members, researchers, people with different familiarity with technology, etc., resulting in a more systematic view in the way technology is conceived and will shape our relationship with the world. The translation of the second workshop participants’ ideas into technical, formal and informal requirements was inspired by this approach, as described in the next sections.

The Application of Imagine to the Project Plan for M³

Imagine makes use of Rich Pictures and EICA as a primary means for data gathering. It was applied in brief (three hour) collaborative workshops run by the team in May and July of 2016.

Research Process – Workshop 1, 19th May

The First Workshop took place on the 19th May in Bletchley, Milton Keynes. 20 participants came to the event and developed their ideas in response to the question:

“What do I Wish I’d Known before I set Out or during My Journey”

Participants were asked to work in three groups of around 6 members and produce Rich Pictures in response to the question (Fig. 5).

Fig. 5

A picture from Workshop 1

The three groups produced a RP each and these were assessed.

Analysis of Results of Workshop 1

The RP produced by each of the three groups was assessed by means of the Content Analysis approach already described. The images were divided up into nine segments and each segment was assessed for content – strong images, visual metaphors, unique ideas, innovative concepts, repeated issues, tasks that needed to be addressed, etc. This assessment of the image was included with a review of the verbal report of each group which was provided at the event.

Figures 6, 7 and 8 are the segmented RPs assessed and Tables 2, 3 and 4 provide the resulting content analysis.

Fig. 6

Group 1 Rich Picture

Fig. 7

Group 2 Rich Picture

Fig. 8

Group 3 Rich Picture

Table 2 Content Analysis of Group 1 Rich Picture

Table 3 Content Analysis of Group 2 Rich Picture

Table 4 Content Analysis of Group 3 Rich Picture

For each group we invited a composer to develop a musical interpretation of the RP. Our intention was to use this as a means to stimulate further creativity in the group at the second workshop.

The composer engaged to provide a musical interpretation of the RP from group 1 suggested the following assessment which can be seen as a companion piece to the Unique Signatory of the RP provided in the Content Analysis:

Group 1 RP is mainly about clear information, honest and integrated. The composer notes:

“I tried to incorporate many contrasting variables that all clearly work together in an organised fashion to reflect an 'integrated transport system'. There is a hint of sentimentality in the harmony which could imply an 'honest' system. The piece has a strong sense of direction musically which could reflect 'clear roads and a clear journey'”.

(A segment of the musical notation for Group 1 is in Appendix 1).

It was observed that the groups listened with interest to the interpretation of each RP as a musical form while gazing the pictures. Group members noted a number of observations about the music. One suggested that it was ‘interesting’ and another noted that it was ‘accurate’. Certainly the music provided a useful ice breaker to the main work of the Workshop process.

The musical assessment of the group 2 RP was as follows:

Group 2 focuses on choices.

“The piece has a slower tempo to imply uncertainty. I have used rising melodies which are allowed to hang in mid-air as a 'questioning' device, and multiple resolutions to the same musical questions in different instruments. All instruments eventually arrive on the same chord at the end so the musical journey is completed safely but with a lot more uncertainty and decision making involved.”

Group 3 the composer suggested were chiefly concerned with Time and Travel:

“This piece has the greatest sense of momentum to convey the sense that time and travel are of the essence. The woodwind part is deliberately busy to convey the 'congested world.' The piece regularly modulates to reflect the 'fractured world.' 'Not all choices are valid' so this piece demonstrates how the right choices will allow the traveller to successfully complete the journey”.

Emerging Themes from Workshop 1

The EICA process provided three strong recurrent themes in the RPs.

• Information:

• All groups discussed the provision of information in some way, and tended to have strong feelings on it. Information came up as something which was unavailable (Group 1: G1), was too poor (G2), provided in excess (G2), or unclear (G3). Groups wanted the right information rather than just more of it. This feeds into choice…

• Choice:

• Did not come up in G1 but did in G2 and G3. Choice was discussed as being limited by various factors, and also having choice was not always regarded in favourable terms (‘decisions, decisions, decisions’) (G2). Choice could be limited by external conditions like congestion, and the weather (discussed by all groups) (G3).

• Groups oscillated between a preference for choice or information. Choice was not necessarily regarded as a useful good because it implies lack of information. When useful information is available, one does not have a choice as such, though this is a positive thing because the information tells you which option is the right option for you (i.e., it is quickest, most affordable, driest etc.).

• Predictability:

• Came up in all groups, and seemed to be the key theme. Groups discussed frustration and anxiety about things going wrong, and the ability of the MK transport network to recover from disruption (G1); the mystery in one’s journey, and the extent to which there are any guarantees the MK transport services will actually provide what they say they are offering (G2 and G3).

This was the key theme. People sought data which would allow them to make a transport decision whose outcome could be predicted.

From these three statements, and working in a similar manner to a Soft Systems development of a Root Definition (see for example the description in Checkland and Poulter 2006), the following “Theme statement” was produced:

People Want Information which Provides Reliable Choice for Predictable Journeys

This synthesis of the users’ wishes, along with the interpretations of the RPs (including the musical interpretation), were to be the main inputs to the second workshop held in July.

Research Process Workshop 2 – 7th July

On the 7th July ten members of the public attended the second workshop. Most of them, 9 out of 10, also participated in the first workshop (Fig. 9).

Fig. 9

Picture from Workshop 2

As Workshop 2 was related to technology design, it was important to build a panorama of the group’s previous experience with technology and their age. Participants completed a form with this information and also stated their consent for having the data collected in the workshop published as a research study.

The age breakdown is shown in the pie chart in Fig. 10, and the relative technology usage made by the participants is shown in the histogram in Fig. 11.

Fig. 10

Pie chart of age of those attending the second workshop

Fig. 11

Histogram of technology usage

Figures 10 and 11 provide us with an interesting observation. Generally speaking those who develop applications for the public are usually young, with a deep and often professional understanding of mobile technologies. In our sample the demographic is elderly and there is a diversity of understanding and experiences with technologies, including smartphones features usage, stated as a not frequent usage for 4 out of 10 participants, and different degrees of familiarity with other transport-related applications, such as SatNavs or Google Maps.

After experiencing the musical signature of the groups, some examples of data available and related technology were briefly presented as a way to inspire participants for the next activities. This presentation included screenshots of different types of GPSs, digital panels, bus-tracker and local weather forecast apps, accessibility resources for visually impaired in bus stops, etc.

Then, two tasks were proposed to the three groups of participants:

• (I) “Describe yourselves using the new technology in your daily life”

A short scenario with a narrative illustrated the expected outcome for this task: “Before leaving home in the morning, Amy checks the BBC travel website. If the motorway (M1) is congested, she takes another secondary route”.

The groups were invited then to share their scenario of technology usage with others.

The results suggested the context in which the technology would be applied by them. Planning a journey in advance, for instance in the beginning of the day, and associating that with other personal daily decisions and commitments were more important than supporting navigation in transit.

The three groups imagined a mobile app, two of them had a complementary website in the scenario.

• (II) “What does your technology look like?”

The second task invited the groups to draft the user interface of their “desirable” technology meeting the challenge of the Theme statement that emerged from Workshop 1. They should evidence what sort of data they expected to see there.

Our literature review informed us that the ways data are employed by users depends on a number of factors, the trade off between their wider ambitions for urban space versus their immediate, personal transport needs; their broader familiarity with public transport networks; and the time at which they seek to access the data available which might be prior to departure or might be en route. In keeping with the openness embedded in the Imagine methodology, our intention was to avoid guiding participants in any way rather than encouraging them towards a specific type of application or data presentation tool. They produced a range of ideas about the potential applications. Some of the resulting ideas are illustrated in Fig. 12.

Fig. 12

Resulting co-design activities

Overall, the three groups suggested a number of concerns, ideas, and wishes in their drafts that were later classified as informal, formal and technical aspects.

The informal aspects raised mainly their perception on current transport-related technology evidencing how they would like to see technology supporting their daily lives, complementing and reinforcing the findings of the first workshop (Table 5)

Table 5 Informal aspects collected in the workshop 2

.

The formal concerns discussed reflected the UK’s transport system organisation, in which private systems work independently, on their own schedules and providing their own online services. Participants recognised this as a challenge when integrating information in a new technology. It is not only about presenting information from different sources together, but also reflecting the impact of one system to the others. For instance, connecting schedules of buses and trains. They also would like to see the system connected to external services that would add extra value, such as buying tickets with a discount or earning loyalty card points.

In technical terms, some participants would like to see the planning system integrated with their personal online calendar. Also, as previously mentioned, they made clear the interest for having a new mobile app, although not all participants have smartphones or use mobile data, considered expensive. A web version to be used at home has been also cogitated, as well as an app using the minimal possible data.

Working from the suggestions contained in groups work, the conceptual designs set out in Figs. 13, 14 and 15 emerged.

Fig. 13

Overview – ‘Shirley’s stress free travel’ (name suggested by one of the groups)

Fig. 14

Making it personal

Fig. 15

Predictability

Discussion

As has already been noted, much work in the area of transport provision in Smart Cities relates to quantitative data focused on expert interpretation of need. The version of the Imagine methodology applied in this case had a clearly participatory focus and provided a small group of stakeholders with the opportunity to engage thoughtfully with the complexities of transport provision. As one attendee at the second workshop noted:

“My company is going crazy trying to plan our transport policy. We thought that this workshop might help”.

Methods to allow citizens to engage in complex planning and assessment processes exist in many domains but are rarely reported with regard to Smart City transport. Despite (or because of) the informal and systemic nature (including sweeping in complexity rather than trying to simplify) of the Imagine method and specifically Rich Picture format seemed to find resonance and even reflief with some of those attending.

The workshop participants expressed their preference for an app that would collect data from different sources, such as traffic reports, weather forecasts, transport timetables and ticket purchase platforms and combine it with user-provided data such as calendar information and journey preferences. One significant element that emerged from the co-design process was the users’ preference for a narrative app interface that could articulate options and provide reminders in a semblance of dialogue with the users. It would be worth pursuing in further research whether this more dialogic tone (“would you like to change your departure time?”) could help app users feel more empowered compared to a more mandatory tone (“change departure time now”). The Rich Pictures from the first workshop depicted users as the passive victims of unknowable events outside their control. The app design process instead told a story casting users as the protagonists of their own travel narrative and making meaningful decisions with the help of data.

One notable aspect of the exercise was the importance attached to uncertainty of information in various forms. Risk, predictability and reliability were all key themes. Participants were keen to assess and reduce the risk of delays and inconvenience in their journeys, they were concerned about the predictability of journey times and the reliability of information from timetables. Those considering walking and cycling were concerned about weather, in which a finite degree of uncertainty is normally assumed inevitable, and often quantified to some degree in forecasts, whereas users of all kinds of motorised transport were concerned about the predictability of delays caused by traffic, accidents, and possibly in some cases lack of coherent transport provision and planning. The extent to which participants recognised the probabilistic nature of uncertainty in all of these sources was unclear and would in any case be variable across a larger user group. Nevertheless, the assumed purpose of transport information to facilitate better choices implies a necessary engagement with decision-making under uncertainty and an information application that failed to deal with this aspect by warning of risks and quantifying likely delays would probably share the same judgement of unreliability conferred on the timetable information available in the present data landscape, that generally ignores this aspect. Consideration and quantification of journey options in a probabilistic sense, perhaps through processing of a growing user-generated dataset, thus appears to be a valuable possibility for future information applications. Anecdotal evidence even points to the potential value of highly specific localised information, for instance regular bus users are sometimes aware of predictable delays to specific buses related to school opening times or regular heavy use by senior-citizen free pass users.

While gathering and analysing such data pose a set of interesting challenges, the presentation of uncertainty information to users provides yet another set of challenges, particularly when information overload has been explicitly cited as a problem. Numerous techniques exist for presenting uncertainty information to non-expert users, such as the use of blur and transparency, cf. the review of Bonneau et al. 2014, and the frequency-based diagrams exemplified in the review by Spiegelhalter et al. (2011). We leave this as an aspect for further research, noting that allowing for uncertainty in suggesting transport options could well lead on to complex decision problems that might also depend on a set of user preferences.

Conclusions

In this paper we have attempted to explain how the rapid and participatory M³ research project has provided innovation in systems method and interpretation of participant engagement. In the specific domain of smart cities research the project has addressed the participation gap identified in the literature by:

• Giving research participants a voice in the data collection landscape, and

• Allowing research participants without technical expertise to have a role in app design

This inclusion provided room for insight and a more holistic interpretation of transport provision and future planning. Further, by providing a musical interpretation of Rich Picture visualisation the project has provided a suggested avenue for further consideration in the encouragement of participant engagement in problem structuring.

Appendix 1. Segment of musical notation Group 1