Panel Session Summary by @rodgerlea

Panelists

• Levent Gurgen. CEA-LETI France.
• Naonori Ueda. NTT Communication Science Labs, Japan.
• Daniel Gooch. Open University, UK.
• Salvatore Longo. NEC Europe Ltd. Germany.
• Masahiro Mochizuki. Ritsumeikan University, Japan.
• Rodger Lea. University of British Columbia, Canada. (Moderator)

Background

As part of the 1 day Smart Cities workshop held at the 2015 UbiComp conference in Osaka, Japan, a panel session was organized to discuss some of the key issues that had been raised in the workshop papers (See: https://www.ht.sfc.keio.ac.jp/iwsc15/program/ ). These issues ranged from the use of data analytics through to business models for Smart City services. A set of pre-prepared questions, summarizing these issues, was sent to all panellists and discussed during the panel session. These key questions were:

• Smart city data analytics – how to unlock the potential?
• Smart city platforms – is there ‘one size fits all’?
• Citizen data – how to gather it, how to protect it?
• City services – how can citizens find and use services?
• Citizen engagement – successful experiences anybody?
• How can we learn from each other’s successes?
• Are there proven business models for Smart City apps?
• What is the killer Smart City App/service
• How do we measure and compare smart cities

While the panel did not cover all these issues in detail, several of the key questions were discussed extensively. This short document attempt to provide a summary of the panel discussion including both the comments/questions amongst the panellists as well as audience questions and comments.

Discussion Points

Smart city data analytics – how to unlock the potential?

Several of the panellists are focused on the analysis of city data with a goal of using advanced algorithms to interpret large volume data sets.  Ueda described some of the work his group are doing in the SODA project and highlighted how time series regression analysis can be used to model and predict city activities, In particular, his group is exploring how to predict and if possible, influence, large crowd movement at events such as concerts, sports games or emergency situations. Longo highlighted the use of a big data platform in the Smart Santander project and his work on preserving privacy while attempting to use personal data for safety applications.

In discussions several date related issues were raised. It was suggested that the larger Smart City projects had moved on from the basic issues of gathering and storing data, and were now focused on using data for analysis and prediction. However, it was observed that although there were several examples of interesting data analysis, as yet, for most Cities, there were very small datasets to work with and it was exceptional for a Smart City project to have access to a wide range of large (depth/time) datasets.

One panellist observed that it still required deep analytical skills to use City data and that what was lacking were simple tools and pre-canned algorithms to help citizens and city managers analyse data sets.

Smart city platforms – is there ‘one size fits all’?

Several of the panellists highlighted their Smart City platform activities. Gurgen highlighted two joint EU-Japan projects. ClouT which has been developing a Cloud based IoT platform for Smart cities and the newer Festival projects, which is exploring federation of testbeds and services for Smart Cities. Lea mentioned the Urban Opus project and its Cloud based Smart City Data Hub that uses Sense Tecnic’s commercial IoT platform, WotKit. Ueda highlighted the Japan based SODA project that is attempting to crowdsource open data and use advanced analytics. Longo highlighted NEC’s focus on re-usable components (Generic Enablers) for Smart City platforms in the context of FI-WARE project.

In discussion, several points were raised. Clearly there are many Smart City projects and many of them have developed their own platforms. It seems unlikely that any one platform, or standard will emerge in the short term.

Smart City projects, whether they are City Scale or smaller scale often focus on different approach or aspects of the smart city. Some are focused on city infrastructure, others on citizen engagement. These are widely different are probably require different platforms.

Interoperability avoids the need to have a ‘one size fits all’ platforms and can be carried out at several levels – by component re-use, at the network or API level, at the data level. It’s not clear today what is actually required for platform interoperability or how to achieve it.

Citizen data – how to gather it, how to protect it?

Many of the projects represented on the panel have a focus on citizen data, ie data gathered from or about citizens as opposed to data about physical city infrastructure. It was clear from many panellists that approaches to gathering and securing citizen data vary widely often strongly influenced by national legislation. Clearly crowdsourcing data was a popular approach, but the issues of privacy and anonymity were always present. Lea highlighted how Urban Opus was attempting to create a non-profit data broker with a trusted service model to give citizens control over use of their data. Gurgen and Longo also highlighted this issue. Mochizuki also described issues with their pedestrian flow and analyzer systems as part of the G-Space project.

There were several issues raised in discussion including the thorny issue of just how much privacy was needed or expected by citizens. It was observed that in some cases citizens seemed quite comfortable in giving up some of their privacy in return for a valuable service – google was cited. However it was noted that citizens are not always clear on exactly how much information they are giving and the ways in which it is being used.

In discussion with the audience later it was observed that many of the researchers felt that privacy was an important issue, but it had come to dominate discussions and was often of less importance to citizens than they (or policy makers) believed.

The issues of trust, anonymization were touched upon by several panellists and it was noted that many of the algorithms under development for city data analysis anonymize as a by product of their approach.

City services – how can citizens find and use services?

This issue was only touched on peripherally. Longo mentioned Smart Santander and its application directory. Lea mentioned the City App Store that Urban Opus provides. Gooch indirectly mentioned that at this stage, many Smart City project were technologically driven and as such, solutions to finding and using services was still in its infancy.

Citizen engagement – successful experiences anybody?

The issue of citizen engagement was a key area for many of the projects. Gooch discussed the MK:Smart (MK:Smart) project and its well developed process for engaging with citizens. Gurgen highlighted activities in both ClouT and Festival that were focused on citizen engagement and indicated that a number of the services had seen end user adoption during trial periods.

There was significant discussion on this issue raising a number of points. Several panellists discussed their experiences with engagement and different approaches. These ranged from the semi formal process Gooch described for Smart MK, through to the incubator model described by Lea for Urban Opus. Both Gurgen and Longo mentioned hackathons and community coding efforts – from discussion it seemed that most projects had used these mechanisms.

While there seemed to be agreement across the panel and attendees about the importance of engaging with citizens, due to the general technical focus of most smart city projects, there are, as yet, no clear examples of how to repeatable achieve this. Equally,  what the engagement results in, and what engagement cannot/should not achieve are still open issues.

Gamification was noted as useful tool for end user engagement but several participants and highlighted as a feature of several service trials.

Overall, it was agreed that engagement of developers and community groups was both possible and practicable, however engagement of large numbers of end users was still an open issue and was tied into service value as well as business models.

How can we learn from each other’s successes?

There was limited discussion of this issue and it was folded into the ‘Compare cities’ question below.

Are there proven business models for Smart City apps? What is the killer Smart City App/Service

Several of the panellists noted applications that had been developed within their projects which had some sort of business drive, however it seemed that most of the project represented on the panel were research driven rather than business driven.

Gooch suggested that the concept of a killer app may not make much sense at the moment due to a) each city have distinct issues and characteristics and b) business and sustainability models are not in place which would lead to an app lifetime to make it “killer”.

Both Uber and Tinder were mentioned as potential killer apps due to their popularity though there is some debate over whether these could be characterised as “smart city”. This suggests that perhaps commercial systems are more likely to result in “killer apps” and that the exploratory research apps/services developed by many of the projects should be treated as laying the groundwork for these commercial services.

How do we measure and compare smart cities

Lea raised the issue of measuring and comparing smart cities and suggested that until we have a good way to compare Smart Cities, it is hard for practitioners to understand what projects do what well, and where to look for experiences and lessons that can be reused.

The panel had a long discussion on the desirability and possibility of comparing across projects and what the KPI’s of a project should be. The panel agreed this was key with some members arguing for directly measurable KPIs leading to a discussion of what the overall aim should be. However, it was also noted by several panellist that different cities have different priorities, eg efficiency of infrastructure, new citizen services, energy, health etc – as such, we need a broad set of metrics or KPIs to capture these different projects.

While there were several opinions on what KPIs could and should be measured, the notion of  “quality of life” as an important concept we should be attempting to improve was raised.

In discussions later in the panel an audience suggestion for an archive of successful projects was discussed. There was general agreement that this would be a useful activity for the community and could be combined with some work to analyse the success stories to develop metrics or KPIs.

Summary

The broad nature of the panel discussion means that a simple summary is not possible, however the following points have some degree of unanimity:

• It is unlikely that any one platform/system will dominate in the short term therefore it is necessary for the community to seek out ways to learn from and incorporate each other’s successes. Some ideas include:
  • sharing components that can be used to build smart city platform
  • sharing applications and trying to agree on some degree of API standardization
  • sharing data by agreeing data formats or catalogues
  • sharing experiences through archiving successes
• Finding ways to measure or characterize different cities and their approaches to developing and supporting Smart Cities apps/services is important and could benefit the community. Several suggestions were made:
  • developing KPIs/metrics for cities and charting these
  • identifying categories of app/service and highlighting successes
  • archiving projects, apps/services and making these easily available to the community
• Engaging citizens in the design, development and eventual use/running of smart city services is critical. However, although there is considerable practical experience in a variety of approaches, no one approach (or set of approaches) can be identified that leads to improved success. The panel highlighted several useful tools:
  • Semi-structured processes incorporating co-design
  • Hackathons and code workshops to engage developers
  • Reaching out to existing community groups to identify needs
  • Gamification as a tool to increase end user adoption

Overall the panel considered a wide range of issues and elicited lively discussion among the panellists and the audience. There was a clear appetite for a more in depth discussion forum and it is obvious that the community has a need for follow-on discussions to explore some of the issues raised during both the panel and the larger workshop.

Acknowledgements

This summary was prepared by Rodger Lea (rodger.lea@gmail.com) with input from all panel members. Additional comments have been provided by Toyokazu Akiyama and Takuro Yonezawa. Thanks to the Smart City workshop organizers as well as the panel organizer Toyokazu Akiyama.