Luke Hespanhol, Martin Tomitsch, Ian McArthur, Joel Fredericks, Ronald Schroeter, Marcus Foth
Governments around the world engage with communities to find ways to better support the interests and concerns of communities and stakeholders, aiming to provide opportunities for more citizens to be involved in decisions that affect them. However, traditional methods of community consultation such as face-to-face meetings and online surveys often fail to reach a representative proportion of the public, as they are not easily accessible, require people to dedicate time and effort, and risk being disconnected from the sociocultural context [1,2].
A number of applications have thus been proposed that allow people—as they are passing through public space—to participate and be more engaged in the discussion of civic topics and city-making. Mobile devices in particular have become a popular choice of platform: They are reasonably ubiquitous in modern urban society, can be used on demand, and allow for more concise expression of opinions compared with conventional written submissions, particularly when leveraging texting and social media . However, those solutions run the risk of excluding whole sections of the community that for various reasons may not own a mobile phone or may not engage with social media on a regular basis. Although mobile solutions are more situated than online forums, they have a number of drawbacks that pose obstacles to community engagement compared with interfaces blended into the urban environment (such as digital kiosks or urban screens). Just like online forums, mobile solutions move the discussion of local community issues to a digital space, which requires community members to first know about the associated mobile channel and second to explicitly access it. This represents a hurdle, and a missed opportunity to make these discussions about shared urban spaces more visible. Overall, these limitations support an argument in favor of platforms that explore opportunistic interactions with members of a local community by physically situating input and feedback mechanisms in or around the public space to which the discussion topic relates.
A range of interfaces has been studied in the past few years, including low-cost interactive posters , gesture-based large projection displays , urban screens , and media facades . A common issue observed in field trials of situated public displays for community engagement is the lack of participation from the public . People usually do not expect public displays to be interactive  and either do not notice the interfaces or feel self-conscious . These effects present barriers to situated community engagement via interactive technologies. Yet previous field trials show that once people overcome these barriers and submit responses, they express feelings of empowerment and connectedness with the local government and broader community. Based on the analysis of relevant works in the field—such as PosterVote , SCSD , and MyPosition —we postulate that to successfully deploy situated polling interfaces, it is crucial to address the following challenges:
- how to increase accessibility to community engagement interfaces, so that a larger section of the community can engage in civic participation
- how to raise awareness about the opportunity to participate in community engagement among passers-by
- how to motivate people to participate
- how to balance visibility of the interface and privacy in the engagement process
- how to provide effective feedback on the interaction with situated interfaces to participants.
In addition to identifying common challenges, our analysis pointed toward some commonly employed strategies when addressing them:
- blending interfaces into the urban built environment for more democratic access
- using public urban screens for real-time feedback on the engagement process
- using tangible user interfaces or full-body interaction as interactive mechanisms and to raise awareness about the interface itself
- ensuring just the right level of playfulness so that an experience is enjoyable yet trustworthy.
The situated character of those interfaces makes them highly dependent on contextual constraints. Nonetheless, there is a lack of comparative studies testing the different strategies within the constraints of the same location and community.
We developed Vote as You Go  to investigate the effectiveness of the strategies listed above. We built two interfaces to deploy in areas of high pedestrian flow in a public space equipped with a large urban screen. We used the screen to create different scenarios that allowed us to study the visibility of the interfaces, the privacy of the voting process, and mechanisms of feedback to participants.
The connection between votes observed on the large screen and the iPad stand as their corresponding input interface was not perceived as obvious.
The first interface consisted of a Web-based survey running on an iPad Air 9.7 inch, installed on a custom stand. This setup provided a tangible point of interaction within the public space, through a physical standalone device with a haptic input mechanism (in this case, a touchscreen). The second interface was a full-body voting application running solely on the large urban screen and using the live footage from a CCTV camera as input. Making use of computer vision techniques, the application then tracked the presence and movements of people in that particular area: Depending on where people stood, the application translated their positions into yes or no votes. If they stood on either area for long enough, the system would compute their votes and restart the process. Therefore, the two interfaces enabled different voting dynamics: While the iPad interface offered a certain level of voter-privacy protection akin to that of other public interfaces such as ATMs, the full-body interface inevitably amplified the participant's opinion to the surrounding public.
Context and goals. Vote as You Go was deployed at the Concourse, a public space in Sydney, Australia, equipped with a large LED screen overlooking a central plaza surrounded by restaurants, a library, and a concert hall. The screen normally features a variety of entertainment content, including cartoons, movies, and documentaries. Our interest was in observing how the different social dynamics prompted by each Vote as You Go interface could affect the levels of participation within that urban context. For that purpose, we structured a series of "in the wild" field studies so we could run the two interfaces with different parameters. We then used those deployments to derive insights about their impact on participation. Common to all scenarios was the location of the interaction zone: a corner in the public space diametrically opposed to the urban screen and continually exposed to pedestrians.
We ran a total of four different scenarios: (1) iPad interface with unrelated content on the urban screen (cartoons, music videos, etc.—part of the regular screen program); (2) iPad interface with the poll results visualization on the urban screen; (3) same as (2) plus the live video camera feed from the interaction zone, each on a section of the screen; and (4) full-body interaction directly with the urban screen (via the live feed captured from its own embedded camera), with no iPad deployed to the public space.
Effectiveness of media visibility strategies on self-initiated participation. In all our scenarios, the great majority of passers-by did not approach the interfaces. That was expected given the casual nature of the voting: We strove to not disrupt the regular crowd dynamics, blending the interfaces into the urban environment. We hoped in this way to prompt citizens with the possibility of expressing opinions quickly, on the go, and, most important, through self-initiated participation. That said, our observational data clearly shows that some scenarios were more successful than others in attracting potential participants and, eventually, leading some of them toward interaction. The version producing the greatest level of participation was the one where the urban screen was partitioned to display both the visualization of the poll results and the live camera feed of the interaction zone. Interestingly, the similar scenario in which the urban screen displayed the poll results but not the live feed produced the smallest awareness levels, even less than providing no feedback whatsoever about the poll on the large screen (i.e., by showing unrelated content).
As we observed, the iPad stand in itself seemed to attract attention by sparking curiosity among passers-by, given its unfamiliarity to the urban space where the study was run. Results displayed on the screen, however, tended to be perceived as a large billboard, and were consequently subject to display blindness: In general, the connection between votes observed on the large screen and the iPad stand as their corresponding input interface was not perceived as obvious. That connection, however, became more apparent when the live footage of the interaction zone—and, consequently, the iPad stand itself—was simultaneously displayed on the large screen, revealing a clear visual connection between what is seen on the large screen and a physical element in the surrounding urban environment. Such a visual connection therefore increased the level of discoverability of the polling interface, leading to a greater level of participation.
The full-body interface produced similar results, although participation itself was more immediate: People needed only to notice the interface while in the footage to prompt it to respond. However, that does not necessarily mean participation was effective or meaningful. Interviews revealed that people were initially attracted by their own reflections on the urban screen. As a participating couple declared, "We were walking along the space when we noticed we were on the screen, so we came back to check it further. We immediately understood how to interact; it was very straightforward." This confirms similar findings in the literature for general public displays , but here the effect was likely amplified by the large scale and highly public nature of the screen, creating for the participants a short moment of fame.
Social interaction and reception by the community. The public screening of the interactive space in the full-body interface also gave rise to collective interaction, thus increasing the number of participants. The full-body interface allowed for groups to dwell in the space for a few moments (while collectively watching the urban screen) and vote simultaneously, a seemingly important requirement for community-engagement interfaces. Collaboration during the voting process itself was also much less common with the iPad: The few occurrences we observed were restricted to social nudging , a voter being told by an acquaintance watching the process what their response should be.
Of course, with both interfaces, it is difficult to tell solely from observations whether participants were expressing their opinion seriously or merely exploring the interface through play. Yet the vast majority of participants we interviewed expressed that their opinions were sincere. Likewise, they revealed concerns about the authorship of the survey and about whether and how the answers they gave would be used. Time to properly reflect upon answers was also seen as a potential issue with the urban screen interface. According to one participant, "I took the questions very seriously, but since I was asked impromptu, I may not have reflected upon my answer as much as I would if I was filling in a written survey." In that regard, the iPad interface seemed to encourage more confident responses.
Despite the somewhat playful aspects of the interfaces, participants seem to have taken them seriously as instruments for community engagement.
The power of amplified mirror images. The integration of urban screens as part of tangible and full-body interfaces (strategies 2 and 3 above) had a great impact on awareness (challenge 2) and participation (challenge 3). Previous works have made use of public screens for two main forms of real-time feedback: displaying visualization of interaction results, and displaying mirror images to reflect the identity of participants and increase their sense of agency [2,5]. For the iPad interface, we tested the former in isolation as well as combined with the latter. While the former type of feedback produced the smallest levels of participation observed, the latter produced the highest. That suggests that combining the display of the poll results with a live display of participants on the large urban screen (as in our third iPad scenario and the full-body interface) is a particularly effective strategy for promoting participation. Although that echoes findings from the literature regarding general full-body interaction with public displays [2,5], we observed that its effectiveness is also verifiable in conjunction with a tangible user interface, here represented by the iPad stand. The iPad stand by itself was not very noticeable. However, when displayed in the large screen alongside the poll results, an obvious visual connection was established between the civic polling and a physical element in the surrounding urban precinct. Such a connection helped communicate to passers-by where to go should they wish to take part in the survey.
The live display of participants on the large screen may have also contributed an element of playfulness and public performance to the otherwise conventional iPad interface (strategy 4 above). Although playfulness did not appear to be a decisive feature in itself (the iPad still attracted some people even when unrelated content was shown on the screen), it was certainly appreciated: The admittedly more playful full-body interface not only yielded the highest participation rates but the interviewed participants also perceived it as highly engaging (challenge 3). At the same time, all interviewed participants said that they quickly learned how to interact with the interfaces, leveraging tacit rules for social interaction: higher degree of collaboration around the full-body interaction; individual voting or social nudging around the iPad stand. Participants also affirmed that they meant the opinions they expressed. In other words, despite the somewhat playful aspects of the interfaces, participants seem to have taken them seriously as instruments for community engagement. The combination of urban screen with either tangible or full-body interaction (strategies 2 and 3) can therefore be seen as reasonably effective when balancing the visibility of the interfaces with the privacy of the engagement process (challenge 4), while providing a good level of feedback to participants (challenge 5).
Design considerations for balancing visibility, privacy, and time for reflection. Our observations point toward a number of aspects that can inform the design of "on the go" polling interfaces for community engagement, notably:
- Blending community engagement interfaces into the built environment (therefore promoting opportunistic interaction) makes them more accessible to the general public, but in itself is not sufficient to grab the attention of passers-by and encourage them to interact.
- Live screening of the interactive space and its resulting playfulness can be an effective strategy for attracting the attention of passers-by and turning them into active participants.
- Public urban screen interfaces increase participation by encouraging group interaction.
- Privately oriented tangible user interfaces (such as the iPad) give people a longer time to reflect upon their answers.
These considerations point toward a need for a hybrid model balanced between private and public aspects of civic participation. As our study suggests, we achieved a possible implementation of this model through the use of the private tangible interface for data entry in combination with the awareness raised by the public urban screen and live screening of the interactive space.
The research presented in this paper was supported by the Willoughby City Council and Urban Screen Productions, and funded through the Henry Halloran Trust.
3. Vlachokyriakos, V., Comber, R., Ladha, K., Taylor, N., Dunphy, P., McCorry, P., and Olivier, P. PosterVote: Expanding the action repertoire for local political activism. Proc. DIS'14. ACM, New York, 2014.
6. Hespanhol, L., Tomitsch, M., McArthur, I., Fredericks, J., Schroeter, R., and Foth, M. Vote As You Go: Blending interfaces for community engagement into the urban space. Proc. C&T'15. ACM, New York, 2015.
Luke Hespanhol is a Ph.D. candidate at the Design Lab at the University of Sydney, researching the potential and applications of interactive urban media architecture. He is also a media artist and lecturer in computational design at UNSW Built Environment, UNSW, Australia. firstname.lastname@example.org
Martin Tomitsch is an associate professor and head of the Design Lab at the University of Sydney, founding member of the Media Architecture Institute, and state co-chair of the Australian Computer-Human Interaction Special Interest Group. His research focuses on interaction design principles for integrating digital technologies into urban space. email@example.com
Ian McArthur is a senior lecturer at UNSW Art & Design. He is a hybrid practitioner working in the domains of speculative multidisciplinary practice, transcultural collaboration, and education change. His research explores urban space via experimental sonifications for responsive media environments. firstname.lastname@example.org
Joel Fredericks is a Ph.D. candidate in the Design Lab at the University of Sydney. His Ph.D. research investigates how urban interaction design and digitally augmented pop-up interventions can enhance community engagement within the built environment. email@example.com
Ronald Schroeter is a postdoctoral research fellow at CARRS-Q, Queensland University of Technology. He developed Discussions in Space, a fast-paced, short-text platform for public discourse using urban screens and mobile phones. firstname.lastname@example.org
Marcus Foth is founder and director of the Urban Informatics Research Lab, research leader of the School of Design, and professor in interactive & visual design at Queensland University of Technology. His research focuses on adopting HCI and design methodologies to build engagement around emerging issues facing our cities. email@example.com
Copyright held by authors. Publication rights licensed to ACM.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2016 ACM, Inc.