12. Rehearsing Emergency Scenarios: Using Space Syntax and Intelligent Mobility Modelling for Scenario Visualisation and Disaster Preparedness

This chapter explores the capacity of two research approaches—space syntax and intelligent mobility modelling (IMM)—to improve the ways people prepare for climate emergency challenges. Wildfires and floods change the environment in unpredictable and catastrophic ways. During these events, the places where people would typically find shelter and the paths they routinely take to get there can no longer be relied upon. Buildings can be destroyed by fire, roads and rail lines can be blocked by flood, and the environment can be obscured by smoke and debris. Both wildfires and floods disrupt people’s understanding of space, undermining their ability to navigate it safely. This is why there is a need to prepare people for dynamically changing environmental conditions to increase survival in the face of climate emergencies. To borrow a concept from the arts, effective preparedness for such a situation requires practice and rehearsal before the event. But how can advanced spatial research support this goal?

Space syntax and IMM are used for a range of purposes, from simulating types of human behaviour to predicting the impact of movement-related decision making. This chapter considers how these approaches can be used to immerse people in spatial data, allowing them to experience, practice, rehearse and perform appropriate responses to wildfires and floods. This goal reflects the growing realisation that the visualisation strategies used in the arts and in design are effective ways of making advances in socio-spatial modelling more accessible, immediate, engaging and compelling for the general public (Gattenhof et al., 2021; Ludlow & Travis, 2018). Proposing an arts-based approach to translating data and methods from space syntax and IMM into an immersive rehearsal space is possible precisely because they have a common foundation in two factors: space and cognition.

The capacity to develop knowledge about an environment, understand it and then use it strategically is a spatio-cognitive skill. A subset of cognitive psychology, spatial cognition encapsulates the ways people acquire, organise and then apply spatial information. As such, spatial cognition shapes the ways people traverse, explore and locate themselves in the world. People use spatial cognition to commute to work, move about their workplace and decide where to meet colleagues or friends. Collectively, this shapes the volume of traffic on a motorway, the number of people using public transport or the ways crowds leave a concert. As such, while it is individuals who employ spatial cognition, the impacts of large-scale spatio-cognitive patterns shape human society.

The combination of space and cognition into a rigorous model is possible because they have measurable or observable characteristics. This means that researchers can model the relationship between environmental characteristics on the one hand and human behaviours and responses on the other. For example, space has both measurable “geography” (location, length, breadth, height) and “topology” (connectedness, accessibility, control) (Hillier & Hanson, 1984). For many decades, empirical researchers have identified patterns connecting human behavioural responses to spatial properties. Some of these patterns relate to movement or wayfinding, which can be statistically generalised to predict large-scale human activities (Lee & Ostwald, 2020). Others relate to perceptual (directionality, intelligibility, complexity) or emotional (security, exposure, confusion) responses to space, which can also be generalised using a range of methods, from statistical correlation (Ostwald & Dawes, 2018) to machine learning (ML) (Jalalian et al., 2011; Wong et al., 2012).

Multiple fields have developed research connecting the spatial properties of environments and human cognition and behaviour using models or simulations. Two of these fields, which have a common interest in movement and navigation patterns, are the focus of the present chapter. However, while space syntax and IMM can simulate human behaviours or transport choices, their conventional applications are often to networks of spaces or streets with known or predictable properties. This raises the question, what about those situations where this is not the case? For example, can they be used to better understand, visualise and prepare people to make spatial decisions about escaping wildfires or floods or evacuating a city shelled by artillery or targeted by drones?

This chapter commences with an overview of the types of thinking and modelling involved in space syntax and IMM and their application in response to dynamic environmental events, from natural disasters to warfare. The section includes several examples of the transformation of spatial data into immersive simulations. The value of these examples is that they suggest a pathway for translating the findings of advanced spatial methods—typically supported by ML and artificial intelligence (AI)—so that they can communicate more directly to people through environmental simulations, gaming and immersive educational experiences.

The four-part process required for this translation to occur involves developing the capacity to (i) picture or visualise extreme events, (ii) narrate or dramatise them, (iii) rehearse or test responses and finally (iv) communicate how we can better prepare for them. This chapter is focused on the third process, rehearsal, where environmental and behavioural data are combined and visualised to prepare people for extreme events. As such, the penultimate section of the chapter explains the idea of a rehearsal and considers how the arts may employ methods such as space syntax and IMM to improve disaster preparedness and resilience in society. This chapter concludes with examples from the authors, which can be developed in future research. Ultimately, this chapter is about how the arts, and in particular modes of practice, rehearsal and performance, can provide a pathway for big spatial data to prepare people for unpredictable climate events.

Both space syntax and IMM view human behaviours as operating in spatial networks. As an example, at a smaller scale, it is possible to create a network diagram, or graph, of the ways rooms in a house are connected. This network, called a “permeability graph”, encapsulates the social patterns of the people who designed and constructed the house. Similarly, the roads that make up the neighbourhood around the house can be converted into a network or graph of the spaces people will move through to go from one location to another. This can be expanded to create a network of all major city roads, rail lines and cycleways. Each of these networks can be modelled mathematically to explore, for example, the extent to which certain street intersections (network nodes) are critical to traffic flow in a city. At a neighbourhood scale, mathematical analysis of the network of spaces can be used to determine the most critical areas for wayfinding, signage or visual landmarks. While these general principles connect space syntax and IMM, there are also significant differences.

Space syntax is the name for a set of analytical techniques initially developed for architecture and urban design in the 1980s (Hillier & Hanson, 1984). Using various mathematical models developed from graph theory, space syntax has been used to create social, movement and spatio-visual simulations that correlate to human phenomena (Ostwald & Dawes, 2018). Such simulations have been used to support the design of buildings and cities and to shape where people are likely to gather, move through or seek landmarks to orientate themselves (Lee & Ostwald, 2020). IMM uses data, often updated in real time or crowd-sourced, and computational simulations to incorporate human factors into transport network planning (Hsu & Wang, 2015; Lewicki et al., 2020). Whereas transport planning has traditionally been concerned with optimal route selection or congestion management, IMM adds data about human behaviours and social and cultural factors that can shape transport decision making (Pribyl et al., 2022). Applications of IMM provide nuanced and rich predictions of human navigational patterns, which in turn shape the way transportation and infrastructure operate.

Both space syntax and IMM use data to support creating or modifying buildings, cities and infrastructure. The findings developed from their use can be visualised in various ways or modified to test alternative scenarios or impacts. While both space syntax and IMM can support dynamic, cyber-physical modelling or digital twins, they tend to be used in fields where the parameters are relatively clearly stable. For example, once constructed, networks of roads are mainly static in space, and the traffic flowing on them, regardless of whether it is moving freely or blocked, falls within a predictable parameter range for IMM. Similarly, while buildings can be renovated and their spaces altered, such actions are rare in their life cycles, and for this reason, space syntax simulations tend to be most useful during the design process. However, there are examples where both space syntax and IMM have been used for less predictable situations.

Space syntax, GIS, transport modelling and network analysis have all been used to identify optimal locations for emergency facilities (Tian et al., 2023; Zhang et al., 2023). Irsyad and Hitoshi (2022), for example, used a combination of observations and space syntax to examine how people make decisions about flood evacuation. With a focus on earthquakes, Tsai and Chang (2023) analyse the effectiveness of space syntax and network analysis methods for developing mass urban evacuation strategies. Significantly, such urban-scale emergency evacuation events have been investigated using IMM. Waller et al. (2023) analyse travel disruption and traffic behaviour strategies resulting from the mass evacuation of Ukraine after the Russian invasion in 2022. They note that not only is there relatively little research available “on travel patterns during such human-driven large-scale (and sustained) disruptive events”, but “studies analysing travel patterns during large-scale conflicts or invasions appear to be exceptionally limited”. To respond to this knowledge gap, Waller et al. (2023) use a methodology combining crowd-sourced data and origin-destination mapping to examine patterns in travel behaviour in selected Ukrainian cities in the month after the start of the invasion. Waller et al.’s (2021) “origin-destination demand and visualisation tool Rapidex, enables the user to download and visualise road networks for any city using a capacity-based modification of OpenStreetMap”. Using pervasive traffic data, they demonstrate how dynamic models of travel disruption and adaptation can be visualised.

A systematic review of applications of space syntax methods in the analysis of complex environments (Iftikhar et al., 2021) initially identified over 4000 articles, of which only nine were about stressful or perplexing spaces and activities akin to those in a disaster situation. One example of this type of research is the work of Lin et al. (2020), which used an immersive virtual reality (VR) simulation to study whether there were cultural differences in crowd behaviour during a fire emergency evacuation in metro stations. They used the Unity3D game engine to convert 3D Studio Max models of metro stations into a format suitable for the simulation. In a study with similar parameters, Wang et al. (2022) examined the impacts of lighting on peoples’ wayfinding skills during an emergency evacuation. Their study also employed an immersive VR experience rendered using Unreal Engine 4, supported by a universal treadmill to enhance the sense of reality. They set up a series of experiments, which simulated an emergency fire evacuation from an underground railway station. Participants were observed, and their paths through space were traced and analysed.

Meng and Zhang (2014) compared people’s experiences of exiting a VR simulation of a hotel room under two conditions. The first group exited in response to a fire alarm but without any visible signs of a fire, while the second group was confronted with a dynamic, virtual fire. By comparing the experiences of the two groups, Meng and Zhang (2014) developed rich data about the impact of a visible threat on wayfinding and behaviour. They found that the second group experienced “significantly higher skin conductivity and heart rate, experienced more stress, took longer time to notice the evacuation signs, had quicker visual search and had a longer escape time to find the exit”. The simulated fire created “higher physiological and psychological stress”, diminishing the capacity of this group to evacuate the burning building safely. Meng and Zhang (2014) concluded that there is a need for education and practice to improve individual and collective preparedness for disaster situations.

Given these examples of immersive experimental environments using gaming engines (Unity and Unreal) and VR, it is not surprising that “serious games” have been designed to study and prepare people for emergencies. For example, the catalyst for the research of Snopková et al. (2022) was the observation that “people tend to [follow] previously-used and known routes (to retrace) rather than follow evacuation signage. This has proven undesirable, even fatal, in emergencies and such [behaviour] calls for a better understanding of the influencing factors”. In emergencies, “decisions are made according to the perceived situation prevailing, the presence of additional visible cues (e.g. fire, smoke), previous experiences, and the actions of other occupants”. Snopková et al. (2022) tested a series of navigation models drawn from space syntax and network analysis, using an immersive VR simulation and an interactive environment scripted using Unity. One of their findings emphasises the capacity of this approach—VR, gaming engine and socio-spatial analysis—for engaging people in deep learning situations.

A recurring argument in the early 2000s was about the value of embedding the arts in a curriculum otherwise dominated by science, technology, engineering and mathematics (STEM). The acronym STEAM has come to reflect the proposition that the arts (along with the humanities and other creative disciplines) have a crucial role in STEM, developing creative insights and communicating STEM messages to the general public, thereby affecting social change. The basis for this proposition, which is repeated in many places, is that the arts “are crucial to facilitating acute and long-term insights into possible social and environmental interactions, impacts, benefits and consequences for our human condition” (de la Garza & Travis, 2018). Ludlow and Travis (2018), for example, identify the critical importance of using the arts to communicate messages about climate change and extreme weather phenomena, and Travis (2018) demonstrates how data about sociopolitical violence is ineffective for developing an understanding of its impact, without rich visualisation of its human causes and effects.

The arts engage people in galleries, libraries, archives, museums and online (Benneworth et al., 2016). As such, it is often argued that the arts have a heightened capacity to communicate with society and shape people’s attitudes and behaviours (Gattenhof et al., 2021). Of equal importance, the arts provide a framework for critically examining and understanding concepts and visualising knowledge in accessible and original ways. In parallel with the increasing awareness of the power of the arts in society, there has been growing recognition that “design thinking” brings new methods to STEM, supporting divergent rather than solely convergent thinking, developing methods to promote innovation and lateral thinking (Lee et al., 2020; Yu et al., 2021). For these reasons, the idea of using arts-based thinking to explore the potential for space syntax and IMM data and methods to be used for disaster preparedness is a natural one. A practical conceptual framework for considering its potential, which appears in several scientific studies referenced previously in this chapter (Meng & Zhang, 2014; Snopková et al., 2022; Wang et al., 2022), combines immersion with the concept of rehearsal.

The dictionary definition of the verb “rehearse” describes the process of practising an activity that will later be undertaken in public. Thus, for example, a person might rehearse a piece of music and later perform it on stage. In the arts, to rehearse is not the same as to practice. The noun “practice” refers to the process of repeating various isolated techniques or skills to such an extent that they become ingrained or second nature. Thus, practice provides the foundation for rehearsal, preparing a person for a successful performance. For example, a singer might repeatedly practise scales for many years (singing ranges of notes in a precise way) but rehearse a particular aria for an opera for a few weeks before the performance. Significantly, whereas both practice and rehearsal seemingly emphasise repetition, a higher level of dynamism is required when it comes to the final performance.

The most effective performers can respond to their contexts, including audience responses, unexpected events and the vagaries of the performance space itself. At its most extreme, this type of dynamic performance takes on another characteristic, “improvisation”, which refers to the capacity to extemporise or expand a performance beyond its initial content. Significantly, Kendra and Wachtendorf (2007) argue that improvisational skills are critical for assisting people to respond to unpredictable and catastrophic events, from terrorism to climate emergencies. In a similar way, Fowkes and Fowkes (2022) call for new modes of rehearsal and collaboration as a means of communicating climate emergencies to a broader world. Barker (2022) notes the role curators play in articulating the types of practice and performance needed for a world facing climate emergencies. When confronted with unpredictable events, these examples treat people’s behaviours as needing practice, rehearsal and improvisation. As research cited previously in this chapter demonstrates, immersive visualisation provides a powerful means of achieving this (Lin et al., 2020; Meng & Zhang, 2014; Wang et al., 2022).

Consider Waller et al.’s (2021) visualisation tool Rapidex. It provides users with an intuitive ability to understand different traffic flows and their ecological impacts (CO₂ emissions), from the scale of the entire cities to individual neighbourhoods and buildings. The Rapidex model for Berlin (Germany), for example, offers an interactive visualisation that can be scaled and viewed in two or three dimensions (Figs. 12.1 and 12.2). Through this process, Berlin is revealed to be a city where 3.7 million inhabitants travel, on average, 6 km or 13 min of travel time per trip while experiencing a moderate level of traffic congestion (congestion index = 2.8). In contrast, Dubai (UAE), with a population of around 3.3 million inhabitants, has a lower average level of congestion (congestion index = 1.54) and an average trip length of approximately 10 km or 8 min of travel time (Figs. 12.3 and 12.4). Waller et al.’s (2021) Rapidex offers a rich tool for interactive visualisations of traffic congestion and environmental impacts across multiple scales.

A 3 D visualization network of Berlin urban traffic with traffic flow patterns. The network is congested at the at the center-left and center-top regions.

An aerial 3 D illustration of Berlin city with high-rise buildings, neighborhoods, and other spaces. The road networks across them are highlighted.

A 3 D visualization network of Dubai urban traffic with traffic flow patterns. The network is smooth, with slight congestion at the center-top regions.

An aerial 3 D illustration of Dubai city with high-rise buildings, smaller buildings, neighborhoods, and other spaces. The curved and straight road networks across them are highlighted.

Ostwald and Dawes (2013) and Ostwald (2014) provide examples of the potential for immersive simulations of space syntax data to support intuitive understanding and rehearsal. In the first of these studies, Ostwald and Dawes (2013) examine the capacity for different types of syntactic maps (space syntax network representations) to capture potential movement patterns in a famous example of Californian modern architecture, Richard Neutra’s Lovell House. Using 3D simulations, along with complex spatial “dissolves”, people can visualise the house first as a habitable environment and then as a complex network of movement paths and optimal places for social interaction or spatial isolation (Figs. 12.5 and 12.6). In a second work, Ostwald (2014) uses a complex mapping of the experience of movement and sound in another of Neutra’s famous designs, the Clark House (Figs. 12.7 and 12.8). This analysis not only provides a comprehensive overview of the mathematical and spatial properties of the design, which are much debated by historians, but also offers the first immersive experience of both the house and its data. Here, people could experience the spaces as they were intended; hear simulations of sound and music, which the design was famous for; and view the mathematical maps of spatial properties and the associated mathematical data, suspended throughout the space (the coloured lines and numbers in Figs. 12.7 and 12.8).

A 3 D illustration of a multi-story house with the top 2 stories having a sloping roof ceiling. A woman with a child is walking in the corridor of one of the stories.

A 3 D illustration of a multi-story house with the top 2 stories having a sloping roof ceiling. The colored lines indicate a set of syntactic navigation data for the house, representing movement paths and optimal places.

An aerial 3 D illustration of a house. It has various connected compartments with an open top. A car is parked in a compartment with partial walls on the left. A table of 8 by 14 presents the data on the right. The colored lines and numbers indicate spatial properties and data within the house.

A 3 D illustration presents an interior view of a house. 2 virtual people stand within a room beside a shelf with a gramophone and other items. One of the people is in mid-speech, gesturing with a hand. The colored lines and numbers indicate spatial properties and data within the house.

These examples from the authors of the present chapter demonstrate the visualisation of complex data in immersive and interactive ways, as well as support for this from AI and ML. When coupled with past research about fire and flood evacuation from buildings, along with the impacts of war on urban evacuation strategies, both space syntax and IMM have considerable potential to help people practise, rehearse and improvise when confronted with climate emergencies. What, then, is required of future research to achieve these goals?

There are two key challenges to translating the insights of space syntax analysis into an environment suitable for supporting disaster preparedness. First, space syntax methods do not, typically at least, respond in real time to dynamically evolving environments, but this may be a necessity for practice and rehearsal in an unpredictable world. Therefore, adaptable wayfinding, without predefined paths, would be a priority for future research, wherein optimal routes to safety can evolve in response to changes in spatial hazards and altered environmental conditions. With the support of AI, socio-spatial data could be developed into an adaptive guidance system, even included in a smartphone app or the next generation of “smart glasses”, to better prepare people and then guide them in real time.

The second challenge is the process of making spatial data visible and meaningful. The examples in this chapter confirm that making data visible is viable, but even immersive spatial simulations are not necessarily intuitive. Gaming headsets and cars with “heads-up” displays are fast becoming ubiquitous and may offer one means of trialling new ways to make data meaningful. Furthermore, multiple examples in this chapter use gaming engines, often in immersive virtual environments, to both test and understand people’s capacity to adapt to spatial risks. These examples could be used as the basis for a process of practice and rehearsal and, with the support of AI, provide the right experimental environment to test new adaptable wayfinding algorithms.

In IMM, a future application of the research in interactive disaster responsiveness would require novel frameworks for decision making as well as for testing protocols needed to precisely define the emergent modelling approaches due to the existing limitations of IMM techniques. It is critical to note that existing IMM approaches, even those which are used to inform governance as well as societal decision making, are highly limited in their behavioural representation of individual-scale characteristics. For instance, even modern IMM approaches generally simplify highly individualised characteristics—such as the level of preparation, resilience and responsiveness—in terms of the quantification of network conditions. Moreover, most standard IMM approaches are tailored to “typical” conditions where estimating the average is the primary goal. Even for evacuation modelling and other disruption scenarios, the focus is commonly on aggregate network-level phenomena. But, by experimenting with the impact on quantifiable decision making within the highly immersive virtual environment, novel IMM solutions can be devised that are more behaviourally realistic. Insights into the individual human response to quantifiable degrees of preparation will shed light on currently unmodelled phenomena within IMM. Further, by enhancing the representation of individual decision making under varying degrees of preparation and rehearsal, the resulting functional tool would provide an entirely new policy sensitivity for IMM solutions. This increased realism enables the required decision support for policy decisions, such as potential support of societal preparedness initiatives.

To rehearse is to prepare for the real event. It requires both knowledge development through practice and experimentation and knowledge application in the final performance, including a capacity to adapt in real time. In this chapter, the concept of rehearsal is used to conceptualise the relationship between emergency scenarios, decision making and the built environment. Importantly, rehearsal does not occur in isolation. It requires a capacity to picture or visualise the setting, and it relies on the ability to construct a narrative and to communicate a message or emotion. While the present chapter conceptualises space syntax and IMM as supporting AI-enabled, data-driven rehearsal, both computational methods are used for visualising environments and the activities that take place in them.

Finally, while space syntax and IMM often address related problems, they each provide unique perspectives. However, further research is required to benefit from their combination within immersive virtual environments. Specifically, space syntax and IMM are currently limited in their capacity to directly address disaster scenarios and prepare society for them. Nevertheless, the use of space syntax and IMM for scenario visualisation and disaster preparedness provides a powerful new tool for (i) exploring best practices for engaging and even educating the public to prepare society for disaster scenarios and (ii) developing novel individual behavioural insights, which will improve the tools society uses to plan for these scenarios. Future research can unlock this potential and its benefits.