ISCN Spotlight: International practitioners' exchange on urban digital twin development

Key takeaways

There is a widespread commitment to open development principles for urban digital twins (UDT): All projects present at the ISCN Spotlight work extensively with open source and open data.
Utrecht and Japan are currently focussing more on a platform approach and open ecosystems for their UDTs; Cape Town, Munich and Wuppertal are developing their 3D models so far more internally and in a use-case-driven approach for their administrations.
So far, reference data, open data from other administrative levels, aerial survey data and point clouds/LIDAR have been prioritised as data sources. Sensor, IoT and cellular data have not yet been much integrated in the projects.
Other data sources have considerable potential. Crowdsourcing data can strengthen citizens' trust in the administration's data work by comparing it with other public data in the Twin. Data sharing through Data Spaces will make it possible to utilise further private data for UDT applications in the future. The city-wide data ecosystem should be constantly considered and expanded in strategy and development.
An initial use case mapping from the projects showed that so far most applications are from the thematic field of urban planning, followed by infrastructure and environmental management. In Japan, there are also already numerous use cases for crisis prevention and response, as well as for citizen participation and for citizen entertainment and tourism purposes.
Following the „Capability Mapping“ scheme of the DIN SPEC 91607 all projects cover already at least in parts the dimensions of “technical integration“, “knowledge generation“ and “decisionmaking“ for UDT. The “operation“ level, i.e. the manual or even automated steering of urban processes through the UDT has yet to be reached.
Multiple operating models and stakeholder structures can be considered for the longterm use and maintenance of UDT and have to be discussed further.

Using data to make processes visible and controllable - nowhere does this seem more obvious and intuitive than in the vision of an urban digital twin (UDT), the digital image of the city. Its development is an international trend topic and at the same time the focal point to which many individual smart city measures converge to or a common project into which they can be integrated.

The ISCN Spotlight brought together international experts from the project level of UDT development for a two-hour exchange. The guests came from Germany (Federal Ministry for Housing Urban Development and Building, Munich, Wuppertal), Japan (Ministry for Land, Infrastructure, Transport and Tourism, MLIT), the Netherlands (Utrecht) and South Africa (Cape Town).

After an initial overview of the similarities and differences between the various approaches by the ISCN (e.g. through Fig. 1), there was a "tour de table" in which each side briefly outlined the status quo of its project as well as current challenges.

Data sources

This was followed by structured discussion points. The participants exchanged views on the data sources and data standards used in their projects and roughly sorted them according to the intensity of use. This revealed that most projects currently rely on combinations of aerial survey data, point clouds/LIDAR, reference data and open data from other administrative levels. Utrecht's project depends less on aerial data and achieves 3D renderings based on the comprehensive, centralised Dutch database PDOK, which includes building IDs, something many other countries do not have. They also reported a dedicated integration of crowdsourced data, i.e. data collected decentrally on behalf of engaged citizens, via the associated EU project Urban Releaf. The assumption is that comparing and supplementing official data with crowdsourcing data can strengthen citizens' trust in public sector data processing.
Satellite data has not yet been widely used to create the UDT, but it has great potential. Although often having longer update intervals, resolution quality is constantly increasing and the usually publicly provided data pool is becoming more comprehensive and can lift some of the burden of data provision for the municipal level. Wuppertal cited openEO Data from the European Space Agency's (ESA) Copernicus project as an example of this.

The integration of IoT, sensor and cellular data has also not yet been fully realised in the current stages of development. In many cases, this data is still only available on the private side (especially cellular data) and can only be made available to cities via better data ecosystems and data markets. Some cities are also planning to install their own dedicated sensor networks that can feed into the UDT. However, the administration and operation of these sources of live data may warrant more discussion on whether it should be within the scope of municipal offices or rather leveraged by other stakeholder groups. For sensitive data, e.g. underground data, discussions showed that there is no way around proper identity management. Utrecht and Amsterdam have already developed FME scripts and dashboards in the joint Netherlands3D project, through which city employees shall be able to request such data in a specified manner. More fundamentally, these questions around a wider set of data sources point to the need for data spaces and their potential for targeted cross-user group data provision and data sharing as an important complementary to the development of urban digital twins.

Data standards

Next to the employed data sources, the participants also briefly discussed respective data standards. The majority of projects in the ISCN Spotlight use the CityGML standard. In the Netherlands the more recent CityJSON standard is also quite common. The trade-off here is often one of the possible depth of detailing and carried information (with advantages for CityGML) versus easier and faster handling in web environments due to fewer directory levels (CityJSON). The latter can be a relevant factor, especially when it comes to accessibility and opening the UDT to a wider ecosystem, as some models to date overwhelm the processing power of many conventional computers in municipal offices or households. However, various programmes now also enable conversions of the formats in both directions, as was also demonstrated in practice in the Peer-Learning between Munich and Utrecht in the #connectedinEurope programme.

And it is this technical detail on data standards and their implications for the "handling" of UDT models for different user groups that marked a good transition for a second discussion point on the envisaged stakeholder structure for UDT projects.

Openness of digital twin ecosystems

Based on a coordinate system from a paper by D'Hauwers, Walravens and Ballon (2021) [1], the projects were asked to categorise their twins according to their current degree of openness in terms of control and development (x-axis) as well as use and accessibility (y-axis) of the UDT (Fig. 2). An indication of openness to a broad ecosystem is, for example, the accessibility of the UDT for any user via web browser interfaces.
Although there was not enough room to analyse the positionings in more detail, different trends did become apparent: Cape Town and Wuppertal can currently best be described as "inside-in" configurations, as their UDT approaches are primarily developed within the municipal administration, but are preparing for opening to a larger stakeholder group or have already been opened up for participatory applications in individual cases.
Munich's profile is currently perhaps the most balanced in terms of the proposed scheme. In principle, the development is anchored and managed within the city administration and the relevant departments. As in Wuppertal, it was emphasised that the development is "use case-driven", i.e. application-oriented. On the other hand, there are already several collaborations and interactions with other stakeholders, such as the Technical University of Munich, start-ups, or the UDT being part of participation formats with citizens or in peer learning and data exchange with Utrecht.

On the other hand, Netherlands3D as used by Utrecht, and the PLATEAU initiative from Japan can be categorised as "outside-out" types. Both projects offer their 3D models to any Internet user for viewing and downloading via a browser. They also allow users to integrate their own private data into the models. Slight nuance could be made by the fact that the "Utrecht approach" (as already mentioned above) is somewhat less demanding in terms of computing power (therefore slightly higher on the y-axis), while the Japanese approach already has a numerically larger and more diverse de facto integration of its project in a larger ecosystem with numerous aligned activities of other stakeholders and start-ups.
It is noticeable that the two UDT projects managed at national level seem to open up to larger ecosystems at an earlier stage, whereas the projects that are still strongly anchored at municipal level focus initially on internal administrative development and applications. (At the same time, Rotterdam's "Open Urban Platform" is an example of a municipal project that is just as openly conceptualized as projects on higher regional levels.)

Further exchange

Several other points of discussion were briefly raised and remain on the agenda for future exchanges, such as a look at long-term stakeholder constellations in the use and operation of the UDT, steps and challenges towards automations, or community building and maintenance challenges for the open code bases of the projects.

[1] D’Hauwers, R., Walravens, N., & Ballon, P. (2021). From an inside-in towards an outside-out urban digital twin: Business models and implementation challenges. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 8, 25-32.