Developing a visual, interactive, web-based tool is a big part of the AquaSavvy project. This tool shows the different water flows that exist in an urban context, supported by local community-driven data platforms. This tool is built by bringing people and technology together, through systems-of-systems data engineering, urban resource flows and participatory planning.
What will the tool look like?
This digital twin concept is human-centred rather than technocentric. This shift alters the perspective from solely considering aspects of risks and technical infrastructure to also encompass nature, human, and spatial aspects relating to synergy and integration with urban design and development.
Imagine a visual canvas – a white board, or a bunch of post-it notes, that has a map with water flows and all sorts of relevant information underneath it. As you draw lines on your canvas, make connections, ask questions, the map moves and alters to incorporate it underneath. When you ask a question about this landscape, the AI assistant provides some information that could help you get to your answer. That’s what we are working towards.
Imagine, as in games when there is a monster nearby and the area gets a red glow, that this landscape can show danger areas, or visualise an emotion map in the same way.
Imagine, as in movies when you know something big is about to happen when the music tells its own story, that you can make this landscape also tell layered stories, perhaps related to the history of a place, or the local band who lives and performs nearby.
Imagine the landscape changing to show the consequences of damming a place up, or digging a trench, and in this way a community can show their local government what they need done, or explain a problem that keeps popping up when it rains. (We don’t know if this is doable yet, and it will have to be low resolution to work interactively, more minecraft than Fortnite. But this is on the wishlist.)
What will the tool do?
The use case of this digital twin concept is to communicate and visualise broad approaches to urban water resource management from a bottom-up perspective. It is intended to incorporate non-traditional knowledge alliances, and as such is envisioned at a different level of technical detail than is typically considered in professional GIS applications. It is more similar to a game world, in that it is intended to be curated, emergent, open-ended and not photo-realistic. One idea, for example, could be how games visualise hints. Critical pragmatism (Forester, 2012) suggests embracing emotions, and the presence of power dynamics during scenario planning workshops. These emotion maps don’t appear in physics simulations, but could be visualised in a tool through game design approaches like lower or higher light intensity, shadowy areas, or red glows.
Further the intended stakeholders is wider than the conventional, to include not only city authorities, urban planners, utility companies, and other stakeholders who will use or contribute to the digital twin, but also not-conventional stakeholders including the general public, and other-than-human actors like biodiversity, ecosystem components, which could be visualised in the game metaphor as non-playing characters (NPC).
The deeper value of the tool, extending its use as a boundary object, is in the facilitation of knowledge clusters, to build networking between organisations (Pego & Lourenço, 2022).
How will the tool work?
The tool is the visual layer of the supporting platform, which consists of several layers, or components.
The visual aspect may be built with javascript, and we’re inspired by Bruno Simon’s website, and Jordan Breton too. Of course, as an academic project maybe we won’t get it this pretty, which is why we are asking for help through the competition.
The important thing to note is that the code that the platform is built from is open, and can be used to create different tools, or different visual applications.
The platform consists of the data integration, the AI module, and the local data management through the community data platforms (CDPs).
Our data integration is analogous to the approach we think could be used to create the Metaverse. Rather than trying to consolidate a bunch of databases, or external services (i.e., wrangle the data from so many different sources into some huge, slow, patchy, complicated mess), our approach integrates bits and pieces across these knowledge infrastructures, to use what you need, when you need it.
Machine learning models, colloquially known as “cooperative AI” can assist decision-making in the domain of water (Wang et.al., 2024). The AI module is not designed as delivering “the truth” or even specific information, but serves as a knowledge base and to reduce misunderstanding among stakeholders. It is a guide on how to go about obtaining the information. It is useful especially when complex tasks need to be decomposed into smaller pieces. For example the AI module may not be able to give the exact regulations for a local area, but may be able to give advice on where to go to look for permission, assistance, the structure of the local government and how to navigate the bureaucracy.
Through the cooperative AI module, the AquaSavvy tool helps to provide insights into the required governance arrangements even in unassisted environments where actors do not have access to the consortium team or facilitation professionals, so that decision-making is supported and legitimised through the involvement of a wide range of stakeholders while being cognisant of the existing policies.
As a tool that is available to anyone, including marginalised groups, this allows participation by stakeholders who may previously not have had real access to means of advocating for justice and equity.
Of course the AI functionality can also be extended. Multi-agent coordination algorithms can then be employed to provide comprehensive model outputs.
Local data management through the community data platforms (CDPs). Our approach takes a more people-centric approach to data generation and big data management. Where the Living Labs generate data, we will be learning and reflecting on better data management practises, from the data point generated to the large scale knowledge infrastructure. This will contribute to data literacy in our Living Labs, but also give us a better understanding of what people struggle with, and what works well for people when we talk about responsible data management.
Also read: The competition showcasing the tool
