Spatial affordances and visitors’ behaviour: A case study in the Design Museum

Abstract

As informal learning venues and anchors of urban public culture, the spatial configurations and exhibition activities of museums shape public engagement and interaction. However, the contribution of spatial configuration to visitor participation remains under-examined in design research. Taking the Design Museum (tDM) in London as a case study, this paper employs a mixed method combining space syntax with non-intrusive behavioural observation to quantify the relationship between spatial affordances and visitor participation. Based on an on-site empirical study, the paper presents how museum configurations influence visitor dwell times and social engagement. Findings indicate that key spatial variables, namely, integration, choice, and visual isovist, have no significant effect on visitor dwell time density or social interaction within exhibition spaces. Instead, exhibit type and seating density are proved to primarily drive visitor behaviour. These findings suggest that museums are not static spaces but dynamic modes of engagement, offering actionable implications for spatial design in urban cultural venues.

Introduction

From ‘exhibition vessels’ into ‘infrastructure venues’ for public engagement

The role of museums has experienced a huge transformation in the past centuries, evolving from private spaces where the wealthy flaunt their collections to public spaces (Fischer, 2009; Redaelli et al., 2025). The term ‘museum’ derives from Latin (Greek: mouseion), originally meaning a temple to the Muses (Pearce, 1993). In its earliest conceptualisation, it was regarded as a temple for entertainment and contemplation (Abt, 2006). Historically, museums serve a dual purpose: as places of contemplation and deep thought, and venues for leisure and entertainment. They were also defined as places of study and repositories of collections with education and research as their basic functions. Although the origin of museums encompassed the notion of a public space, their initial function was limited to collecting artworks that suitable for the palaces and estates of the nobility (Alexander et al., 2017). This phenomenon began to change towards the end of the 17th century when the University of Oxford established its first museum of a public nature. The conceptualisation of ‘public museum’ has raised. However, access to museums was reserved for the privileged elite in the mid-18th century, whilst the rules and regulations governing visits served to separate the bourgeoisie from the general public (Bennett, 1995). It was not until the emergence of ‘open-air museums’ in the 19th century and ‘community museums’ in the 20th century, which adopted accessible formats closely tied to the everyday lives of ordinary people (Mohamed, 2022; Williams-Davies, 2009), that museums began to embrace the public, towards social inclusion.

The 1990s marked a critical transitional period of museums. In 1995, the International Council of Museums (ICOM) defined museums as non-profit and permanent institutions that provide services open to the public, with a mission to acquire, protect, exchange, and exhibit the material evidence of humankind and the environment for research, education, and enjoyment. The nature of museums has changed as the concept of ‘social exclusion’ has become more widely used in British and European politics, which defined in terms of four dimensions: individuals could be socially excluded from (i) civic integration, (ii) the labour market, (iii) welfare state provision, and (iv) family and community (Evans, 1998). Since New Labor came to power, the cultural sector faced new demands to address inequality and poverty. As a result, museums were required to demonstrate their ability to promote social inclusion, address poverty and disadvantages, and attract a wider audience in addition to their value as educational institutions (Davis, 2007).

In the 20th century, museums gradually emphasised public engagement as a mainstream form of entertainment. When the century began, the concept of ‘ummuseum’ emerged, emphasising museums’ pleasant nature to the public. Towards the end of the 20th century, museums increasingly emphasised experiential and storytelling attributes, embodying social context, and moving towards community engagement and diversity (Alexander et al., 2017). This shift revealed that museums have become places that not only attract visitors but also encourage them to engage repeatedly (Everett and Barrett, 2009; Fischer, 2009).

The transformation in the nature and concept of museums has been accompanied by evolving demands for spatial layout. In recent decades, enhancing visitor experience has become central focus of museum theory and practice. Scholars have emphasised and affirmed the educational and research value of museums as public cultural institutions (Alexander et al., 2017). Studies have proposed that learning constitutes identity formation, indicating that audience acquisition within exhibition spaces comprises multiple factors, including personal background (prior knowledge, experience, and interests), the physical environment (exhibitions, activities, artefacts, and labels encountered), and the socio-cultural environment (group dynamics, interactions between audiences, and the cultural experiential value derived within the museum) (Falk and Dierking, 2000). This implies that when visitor needs align with the affordance of museum functions, a positive experience is generated (Falk, 2011; Falk and Dierking, 2000). Consequently, the mechanisms linking spatial affordances and visitors’ behaviour needs to be examined.

Our investigation is based on the theory of ‘spatial affordance’, originally proposed by Gibson in 1977, emphasising an active exploration for visitors to investigate environment (Gibson, 1977). The term of ‘affordance’ was first described as the ‘fit’ between individuals and their surroundings that creates opportunities for action, whether those opportunities were good or bad (Gibson, 2014). This concept serves as a key mechanism for triggering understanding and action in others, with an emphasis of the interaction between the environment and the perceiver, favouring the study of environment-perception relationships (Tröndle, 2014). In this context, a museum is not a fixed container but a dynamic environment where people, space and exhibits interact to create social dialogues (Thomas, 2010). It is also regarded as an inclusive dimension of museum (Barrett, 2010).

To explore the relationship between spatial affordances and behaviours of visitors, this paper examines an integrated analysis on how spatial configuration and visibility relate to visitor coexistence patterns, with a focus on the Design Museum (tDM). As the world’s first museum dedicated to design, tDM employs its location, the building itself, and its surrounding environment to enhance public understanding of design through multiple dimensions (Lu, 2012). The analytical focus shifts from outcomes to processes, primarily examining how spaces induce specific behaviours. Two research questions will be addressed:

• RQ1:

How do spatial configurations and visibility influence visitors’ dwell times, encounters, and social interactions?

RQ2:

Which spatial combinations most facilitate social engagement of visitors?

To address these questions, we employ a mixed method, combing space syntax to quantify the selected case study with non-interventional behavioural observation to analyse how integration, choice, and visual isovist within the permanent and special exhibitions correlate with visitor movements. Based on the findings, we propose key indicators for spatial affordance to promote social participation, form a ‘spatial affordance-behaviour’ framework, and offer design implications for future museums.

Literature review

From Gibson’s affordance to spatial affordances within the museum context

The theory of affordance mainly discusses the transactions between human behaviour and space (Gibson, 1977). Gibson suggested that active exploration was the key concept of the theory (Gibson, 2014). His view prefers to focus on the environment itself, especially the everyday environment, over the observer, which has generated a profound impact on the field of architecture. He insisted that perception did not come from the perception of basic architectural volumes. Most people do not see form and shape when they look at places; they see the possibilities that the environment offers for people’s activities, that is, what the place can do for them (Greeno, 1994). Affordances of the environment refers to the possibilities offered by the environment as people move through a space, perceive objects and act in response to situational needs. It exerts a powerful guiding influence on spatial construction (Tzortzi, 2015), and can be specifically divided into two dimensions: the physical environment and the embodied experience of human beings. According to Gibson, architecture should return to the essence of everyday function rather than formal art, suggesting that the environment provides a variety of functions which offer possibilities for a variety of people’s activities, and that perception is the direct awareness of these possibilities. Thus, the Theory of Affordances also emphasises the purpose and function of perception.

It has been acknowledged by Bonnes and Secchiaroli (1995), who further indicated that physical environment could restrict or support human behaviour. These affordances would impact the possibilities of actions by cues in the environments, and body would react based on its surroundings (Franck and Lepori, 2000). Based on that, Atmodiwirjo enriched the theory and applied it to explore how spatial affordances influenced the sensory integration of autistic children. The results showed three physical levels, including spatial property, positioning and complementation with other elements (Atmodiwirjo, 2014). Although the study focused on a specific user group, it provided a transferable analytical framework for understanding how a space can guide or inhibit particular behaviours by decomposing affordance into concrete spatial dimensions. Tzortzi then expanded it from a spatial perspective by analysing the correlation between spatial layout and embodied experience, which advocated that space and time should be considered to discuss the complex of experiences (Tzortzi, 2017).

In the museum context, the term of affordance indicates possible behavioural experiences resulting from the influence of the surrounding environment (Achiam et al., 2014; Peponis and Wineman, 2002; Tröndle, 2014; Wineman and Peponis, 2010). Wineman and Peponis (2010) discussed the spatial affordances in museums from the perspectives of accessibility, visibility, and visitor movement. Informal education in museums is defined as a moveable spatial behaviour, in which the educational information is constructed and conveyed through movement in space. Spatial layout also affects the socialisation of knowledge and the ‘how’ rather than the ‘what’ visitors learn (Peponis, 1982). In this sense, spatial layout generates different interactive, self-contained units of knowledge from the exhibition content (Peponis and Wineman, 2002). By regulating the visibility patterns of the space, multiple routes and conscious choices are automatically generated between exhibition elements. Visitors’ choices directly influence the structure of the knowledge they acquire in the exhibition.

Tröndle (2014) acknowledged the influence of spatially guided movement by investigating how the architectural and curatorial environment affected visitors’ attention to various artworks, as well as how visitors moved through the facility. The results emphasised that curators should pay greater attention to the organisation of visitors’ spatial behaviour in order to meet their need to maintain an appropriate distance from, or proximity to, each artwork. In this context, the museum is considered as a force field of external factors including single artworks, groupings of works (e.g. series or compositional hangings), socialisation, and individual spaces, etc. This interwoven space influences the consciousness and behaviour of museum visitors.

Furthermore, the concept of affordance is also reflected in the design of exhibits within the museum context. Norman developed the theory from a design perspective, focusing primarily on perceived affordances rather than true ‘physical’ affordances (Norman, 1999), emphasising the implicit use of perceived affordances, which has been applied as inspiration for interactive exhibit and interface design. The study of Achiam et al. (2014) echoes Norman by explaining how differences in exhibit characteristics lead to differences in the interpretive strategies that visitors use in their meaning-making process. This argument elaborates that actions do not necessarily produce corresponding perceptions, and excessive gaps between exhibition intent and visitor experience may result in exclusion, which highlights the importance of coordinating spatial and interpretive affordance to minimise exclusionary effects and broaden participation.

Museums and social engagement

The public nature of museums remains widely debate in existing literature. Habermas’s concept defines universal and inclusive access as the features of public space (Habermas et al., 1974). In terms of ownership, many museums are privately owned, creating barriers to entry, even some of which are without entry fee (Barrett, 2010; Staeheli and Mitchell, 2007). Zukin (1995) further argues that public spaces can promote cultural construction in cities, yet museums remain intrinsically linked to economic and private interests. However, many scholars recognise museums as public spaces from the perspectives of function and mechanism for public engagement.

Museums serve the broader public interest as cultural and educational institutions, even some of which are private ownership (Weil, 1996). With the first public museums being born in universities (Alexander et al., 2017), the educational role of museums was inherently shaped by their institutional roots in universities. For example, students study specialised museum courses to acquire skills for more diverse jobs (Johnson et al., 2017). Thus, another dimension of museum education lies in promoting employment, which demonstrates the social nature of museums.

Beyond education, the concept of museums as public institutions centres on their roles as spaces within the public sphere for constructing shared culture and collective memory, referred to as a civic laboratory (Barrett, 2010; Bennett, 2005). This democracy manifests both in public participation and in the public’s right to voice their opinions (Barrett, 2010). Visitors naturally consider that museum is a physical space and they can interact with exhibits, and sociability is regarded as a defining element (Staeheli and Mitchell, 2007). The concept of the participatory museum proposed by Simon further advances this idea, framing museums as connection to facilitate interaction between communities and visitors, thereby enriching the definition of museums beyond exhibition spaces (Sandvik, 2011). From this perspective, participatory museums transcend spatial boundaries to function as catalysts for social interaction. In the case of the Design Museum, this participatory vision is realised in the permanent exhibition Designer Maker User, which features everyday objects donated by visitors, demonstrating the shift of visitors from passive audiences into participants in the exhibition narratives.

Existing research has further explored the social engagement of museums by examining visitors’ motivations and types of participation. Falk categorised visitors based on their identity motivations (Falk, 2011), while Pruulmann-Vengerfeldt and Runnel classified visitors according to their modes of museum engagement (Pruulmann-Vengerfeldt and Runnel, 2014). The museum is considered as a zone of communication connecting community and visitors (Vårheim and Skare, 2022). Redaelli et al. (2025) highlights the role of museum as the intersection of the public sphere and urban design in fostering social interaction and inclusion. These findings demonstrate that museums, as public spaces, extend beyond their physical dimensions and reach into social and community sphere, which also serves as a response to the idea of ‘museum without walls’, first articulated Malraux (Barrett, 2010; Malraux, 1967). Integrating visitors’ behaviour, space and participation, the mechanism between museums and social engagement can be analysed from both temporal and spatial perspectives (Wang and Huang, 2024). However, there remains a research gap regarding the quantitative analysis of how the spatial affordances of museum spaces support social engagement.

How space syntax indicators relate movement and dwelling

Space syntax is a theoretical framework proposed by Bill Hillier in 1970s (Hillier et al., 1976) for the systematic analysis of spatial configurations. It is usually applied to analyse the relationships between spatial layouts and social interactions within geographic and urban fields. Based on its definition, this method treats space as a system, focusing on the integration of space rather than measurement (Wineman and Peponis, 2010). Using visibility graph analysis (VGA), the primary objective of space syntax is to assess the relationship between spatial layouts and visitor movement patterns. Three indicators can be used to measure the layout, namely, integration, choice, and visual isovist (Hillier and Hanson, 1984). Integration indicates the accessibility from one space to another, and choice means the shortest distance from one by one or the probability that a certain axis or street is traversed. It means that the accessibility of a space can be measured by the two indicators. An isovist refers to the set of all points visible from a specific observation point in space relative to a particular environment (Benedikt, 1979), and is commonly used to measure the visibility in space (Hillier et al., 1987). Gibson (2014) has noted in early years that isovists are capable to capture objective properties.

This method has been widely used to evaluate the relationship between visitors and spatial affordances in the context of museum. Hiller and Tzortzi (2006) utilised the methods of space syntax to explain how the spatial construction of a museum influenced visitors’ movement, including integration, connectivity and visibility. However, the study focused on the analysis of geometry and topology, yet failed to adequately consider the influence of content and media. Centred on the theory of affordance, Wineman and Peponis (2010) conducted a study employing methods such as field observation, path tracing, and dwell behaviour recording to analyse how spatial factors, including integration, choice, and field of view, facilitate or constrain visitors’ exhibition-viewing behaviours. It has been acknowledged by Tzortzi who argued that experience of visitors was the reaction of different types of spatial design, and proposed that embodied experience of museum would be created by offering three key elements of spatial affordances: integration, choice, visual isovist (Tzortzi, 2015, 2017).

Tzortzi (2015) extended the research scope by proposing the idea of spatialised curation, emphasising the integration of display with space syntax to influence visitors’ cognitive pathways, thereby enhancing the consistency of the guided experience. This has enriched the theoretical context of space syntax, and highlighted the interconnection between spatial layout and curatorial narratives. Additionally, space syntax has been applied to describe human’s embodied experience in museums by analysing pathway design, resting areas and sightlines within the museum (Tzortzi, 2017). Four spaces, including occupation space (a), control space (b), circulation space (c) and choice space (d), are used to describe the spatial layout (Hillier, 1996; Tzortzi, 2017). The spatial sequence in museums is primarily guided by (c), supplemented by (d) to offer multiple route options, with (a) and (b) occasionally employed (Hillier and Tzortzi, 2006). These elements are combined to direct visitor flow. However, it relies on qualitative descriptions and case-based reasoning, with limited quantification.

Understanding spatial affordances for cultural spaces

In summary, extensive research has established that space syntax provides an effective method for analysing the accessibility of physical spaces and buildings (Hillier, 1996; Teklenburg et al., 1993; Tzortzi, 2015; Wineman and Peponis, 2010). Tzortzi (2017) addresses a limitation in existing studies by emphasising the interplay between spatial layout and display in shaping visitors’ embodied experiences. This insight has motivated our research to explore two layers: the physical space and proactive initiatives, with an emphasis on the relationship between visitor behaviour and available facilities. However, most studies mainly focus on the interior of museum spaces, with limited analysis of content and related factors.

Furthermore, space syntax has inherent limitations, such as insufficient consideration to three-dimensional aspects like building heights and wayfinding (Yamu et al., 2021), as well as a failure to adequately capture the influence of exhibition types on behaviour (Dai et al., 2025). Consequently, additional research methods are required to thoroughly analyse the relationship between human behaviour and spatial configurations. Considering the definition of museums as public spaces, their role as social facilitators extends beyond mere physical dimensions, indicating the need to investigate their dynamic role as cultural disseminators. Further exploration is needed to understand how spatial configurations influence visitor dwell times and social engagement, thereby enriching the understanding of spatial affordances for cultural spaces.

Methodology

Case and context

The Design Museum in London is the world’s first museum dedicated to design, with a mission to enhance public understanding of design. Relocated in November 2016 from the Butler’s Wharf warehouse to the Grade II listed former Commonwealth College building, the museum consistently adheres the principle of sustainable reuse of historic buildings (McDermott, 1997; Wilson, 2016), which facilitates the repurposing of existing building structures, and fosters discussion on the transformation of historical spaces and their connections to visitor behaviour (Figure 1). Exhibitions are generally divided into permanent and temporary displays. We select two exhibitions – a temporary ticketed exhibition and a permanent exhibition, presenting distinct spatial typology: liner and dispersed formats (Austin, 2020). This aims to explore the relationship between different exhibition layouts and visitor behaviour.

Figure 1.

Design Museum (Old venue-before 2016 and new venue-since 2016) (The authors).

The exemplary geographical positioning of the Design Museum’s new London site manifests in three key aspects: transport accessibility, cultural clustering, and venue openness. Firstly, the exhibition space lies approximately 600 m in a straight line from the nearest underground hub, High Street Kensington (served by the Circle and District Lines), a journey of roughly 10 minutes on foot. Secondly, the south-east flank of Holland Park borders South Kensington, an area with major national museums such as the Victoria and Albert (V&A) Museum and the Natural History Museum. The new museum’s establishment here is therefore entirely logical. Thirdly, it shares an unobstructed boundary with Holland Park, allowing local residents to visit the museum while strolling through the park. Similarly, museum visitors can directly access Holland Park for walking.

Methods

This study employs participatory observation with a mixed-methods approach. No personal data is collected, and all photographic material is de-identified for aggregate analysis only. Space syntax analysis was conducted on the Design Museum’s permanent and temporary exhibitions, while qualitative research was applied to other public and front-of-house areas. Human behaviour was observed and documented through photographic recording and coding. The field observation of the study was conducted over one working day from 10 am to 4.30 pm, covering the museum’s regular opening hours and midday period. We documented visitor behaviour within two exhibition galleries (the permanent and the ticketed temporary ones) and public spaces – such as the café and gift shop – through photographic recording. Visitor behaviour data was gathered through non-interventional behavioural observation, and all photographs were taken in a de-identified manner without contact or interference with visitors.

Given the high visitor mobility in museum environments and as this study aims to analyse how spatial configurations influence visitor dwell time and behaviour, researchers tracked visitors within the observation area during fixed time intervals, recording those within their line of sight (Kawulich, 2005). Visitors behaved naturally without prior recruitment or experimental intervention. Totally 77 selected visitors were discreetly tracked. We primarily employed quantitative methods to calculate micro-environmental instances of spatial affordances, such as seating density and the relationship between exhibits and visitor behaviour. Behavioural patterns extracted from observational data included visitor dwell times and analysis of social interactions.

Data analysis

Space syntax was primarily employed to analyse the configuration of the two exhibition halls, including integration, choice, and visibility graph analysis (VGA). Observational methods mainly recorded visitor dwell times, seating density, and exhibit types (dynamic and static) to analyse their correlations. Dwell time was primarily measured visually by on-site researchers. Photographs were used to verify the number of people, gathering locations, and their behaviours. Linear regression analysis and correlation analysis were employed to examine the correlations between various indicators and their degree of influence (Dao-de, 2000). QGIS and Depthmap were used to visualise spatial configuration in the museum. Given the location in the United Kingdom, we employed the British National Grid (EPSG: 27700) – the most commonly used coordinate system for processing UK geographic data within QGIS (EPSG database, 2021), to analyse the two floor plans.

The entire area is broadly divided into three sections: the forecourt (connecting the venue to the street), exhibition halls and public spaces within the building, and the café and shop. The external plaza links to the neighbouring residential area, attracting many residents to have a rest or engage in recreation. Outdoor features include a pool and seating for public recreation and relaxation (Figure 2). It illustrates footfall comparisons in the forecourt at 10.30 am and 4.30 pm local time, revealing higher density during the afternoon period. The entire museum comprises five levels (Table 1). Upon entering the main gate, visitors first encounter Level G, housing temporary exhibitions, the café, and the shop. Level 1 houses studios and a library, with corridors also serving as temporary exhibition spaces. Visitors pass through these areas when moving from Level G to Levels 1 and 2. Level 2 features permanent exhibitions and a restaurant. Basement Levels B1 and B2 comprise additional exhibition halls and storage rooms. This study primarily analyses the special exhibitions (ticketed) on Level G, and the permanent exhibitions on Level 2.

Figure 2.

Population density comparison of morning and afternoon at the Design Museum Plaza (New venue) (The authors).

Temporary exhibition (admission fee)

Upon entering the main gate, the space directly ahead serves as the temporary exhibition hall. Seven visitors were present. When visitors approach individual exhibits sufficiently closely, they could fully comprehend the visual content of each piece and are most likely to engage with the exhibits (Wineman and Peponis, 2010). Therefore, we consider that when visitors linger before a particular exhibit element, an engagement would be registered, indicating that this exhibit has the potential to provide an intensity of ‘stay’. Using the Depthmap and QGIS, we first analysed the integration, choice and isovist within the exhibition hall. Figure 3 describes the basic layout of the exhibition hall, including the walls, traffic flow, entrance, exit, and other elements descriptions. Figure 4 depicts the visibility graph analysis (VGA) of Figure 3. The background colour bands indicate the accessibility ranging from low (cool-cyan) to high (warm-red), progressing from cyan → green → yellow → red, signifying gradually increasing integration. The central serrated main corridor exhibits a continuous yellow-to-red hue, forming the visual focal point of the entire space. Peripheral rooms and terminal corridors predominantly display green-to-blue tones, constituting ‘deep/concealed’ zones. Multiple doorways and turning points form warm-coloured ‘nodes’, representing potential high-traffic congestion points. To further validate this, we conducted a correlation analysis on integration, choice, and intensity (Table 2). Pearson’s correlation coefficient was employed to quantify the strength of these relationships. Analysis reveals a correlation coefficient of 0.188 between choice and integration, significant at the 0.01 level, indicating a strong positive correlation. The correlation coefficient between choice and intensity was 0.105, significant at the 0.01 level, confirming a significant positive correlation between choice and intensity. The data analysis results corroborated the visualisation findings.

Figure 3.

Floor plan of temporary exhibition hall (The authors).

Figure 4.

VGA of temporary exhibition (The authors).

The isovist area (the area visible at a glance from a given point) of the layout was analysed. The principles for selecting sampling points are primarily based on: (a) significant turning points and nodes, and (b) entrances and junctions of different exhibition zones (Wineman and Peponis, 2010), to reflect the heterogeneity of spatial visibility and orientation (Figure 5). We designed the visitor flow primarily around spatial transition points as given points. The average isovist area for the 19 sample points is approximately 36.1 units. High-value isovists are typically located in the centre of the exhibition hall, while low-value ones are generally found in corners or small partitions (Table 3). In Figure 5, the orange circles represent 19 sample points. The larger the circle, the greater the isovist area it covers; conversely, the smaller the circle, the lesser the area it encompasses. Findings indicate interactive exhibits predominantly occupy areas of high accessibility, with seating typically positioned at inflection points. Furthermore, zones of high visibility largely coincide with regions of high integration, demonstrating the exhibition hall possesses dual characteristics of high visibility and centrality. It should be noted that points with low visibility are generally distributed at turning points and angular areas. The high choice density observed at the turning point on the left side of the plan indicates these areas present congestion points for pedestrian flow. Accordingly, it is necessary to pay attention to crowd congestion and evacuation pathways. Additionally, it can be noted that areas with high population density tend to be concentrated in locations with relatively high choice and integration.

Figure 5.

Visualisation of VGA of temporary exhibition (The authors).

Then we analysed spatial layout to determine whether there was a relationship between exhibit and seating density and visitor behaviour. We mapped visitors’ dwell times to exhibition hall coordinates on a floor plan. As shown in Figure 6, after hierarchical colouring, we observed that visitors lingered longer in areas with higher seating density. We therefore hypothesised that seating density influences dwell time. The prevalence of interactive exhibits surrounding seating areas may also indicate that such exhibits influence visitor dwell time, suggesting a potential correlation between seating location and exhibit category. Consequently, we conducted a correlation analysis (Tables 4 and 5) to examine the relationship between seating, exhibit category, and visitor dwell time. Locations with seating were coded as 1, while those without seating were coded as 0; static exhibits (paintings, static installations) were coded as 1, and interactive exhibits (exhibition booklets, moving images with/without sound, headphones, touchable exhibits) as 2. For dwell duration, we coded durations as follows: 1–5 minutes as Group 1, 5–10 minutes as Group 2, 10–15 minutes as Group 3, and 15–20 minutes as Group 4. Results indicate a correlation between seating and exhibit type, suggesting interactive exhibits are typically situated in areas of high seating density.

Figure 6.

Dwell gradient of visitors (The authors).

However, neither seating density nor exhibit type correlates with visitor dwell duration. To assess potential social engagement, we analysed co-presence existed within the gallery using Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering on the crowd flow data (Ester et al., 1996). The neighbourhood radius was set to 1.5 m, a threshold aligned with the standard social distance defined in sociological research (Goldstein et al., 2023). The results indicate that no clusters were generated, revealing no detectable sustained spatial co-presence between visitors in the temporary exhibition hall (Table 6).

Permanent exhibition

Employing the same methodology, we analysed the permanent exhibition Designer Maker User at the Design Museum. The results proved entirely different. The numbers of visitors here were dozens of times higher than those for temporary exhibitions. This exhibition selected over 1000 classic designs from the 20th century to the early 21st century that profoundly influenced contemporary life. Its curatorial approach eschewed traditional chronological layout, instead exploring the intersection of design’s impact on popular culture, technological change, and consumer-driven environments (Wilson, 2016). The exhibition layout is divided into three sections (Figure 7). We recorded approximately 70 visitors. To ensure accuracy, we utilised Python to detect the number of individuals in photographs (Figure 8). We also aligned visitor coordinates with the exhibition floor plan for cross-validation. Due to the use of non-interventional observation, we simplified behavioural distinctions into viewing exhibits and not viewing exhibits. Exhibits were categorised into static displays, including paintings and texts, and dynamic displays, encompassing video installations and those requiring visitor interaction (Lu and Chang, 2015).

Figure 7.

Exhibition zone layout of permanent exhibition (The authors).

Figure 8.

Person-detection (The authors).

We conducted a correlation analysis to examine the relationships between the exhibition hall’s integration, choice and intensity metrics, with results presented in Table 7. The correlation coefficient between choice and integration was −0.070, with a p value of 0.496 > 0.05, indicating that no significant correlation existed between these two indicators. The findings mean that the spatial integration of the exhibition hall was unrelated to the planning of moving routes. In contrast, the correlation coefficient between choice and intensity was −0.447, exhibiting significance at the 0.01 level, demonstrating a significant negative correlation between choice and intensity. Based on that, it can be concluded that integrated layout of the exhibition space facilitates increased visitor flow, whereas an excessive number of circulation routes diminishes visitor dwell intensity.

Table 8 presents the results of visual isovist analysis for this exhibition. Points with larger isovist areas offers a wider field of view, tending to form ‘visual convergence/pausing’ zones that can improve spatial legibility and attract visitor attention, such as exhibition focal points or gallery transition nodes. Reference points (Refs) 1, 2, and 8 were regarded as potential visual nodes based on their isovist values. Points with smaller isovist areas were often situated in enclosed or perimeter zones (Ref 0, 3, 4), where visual flow and circulation are obstructed, leading to constrained movement paths or shorter dwell time.

Figure 9 indicates that the isovist areas mapping for the three exhibition modules of the permanent exhibition are relatively balanced. Overall, the lower-middle and left-side areas of the gallery offer a comparatively broader visual field. Building on this spatial analysis, we conducted a correlation analysis to examine the relationship between seating, exhibit layout, and visitor dwell time. The visualisation reveals minimal overlap between clusters (identified via DBSCAN) and integration scores, yet strong overlap with choice scores. Specifically, the largest cluster was primarily distributed around the exhibition hall entrance, while visitor distribution was relatively even in other areas. This finding suggests a high level of spatial co-presence and population density at the exhibition entry point.

Figure 9.

VGA of the permanent exhibition (The authors).

Table 9 indicates a linear regression analysis conducted with exhibits, seating, and behaviour as independent variables and dwell time as the dependent variable. The coefficient of determination R² was 0.833, indicating that exhibits, seating, and behaviour collectively explained 83.3% of the variance in dwell time. This model has passed the F-test (F = 109.942, p = 0.000 < 0.05), indicating that at least one of exhibit, seating, or behaviour influences dwell time. Detail analysis reveals that the regression coefficient for exhibits is 1.101 (t = 6.270, p = 0.000 < 0.01), indicating exhibits have a significant positive effect on dwell time. The regression coefficient for seating was 0.551 (t = 3.328, p = 0.001 < 0.01), indicating a significant positive effect of seating on dwell time. The regression coefficient for behaviour was −0.050 (t = −0.307, p = 0.760 > 0.05), indicating that behaviour had no significant effect on dwell time. In summary, exhibits and seating have a significant positive effect on dwell time, while behaviour has no significant effect.

We further conducted a linear regression analysis to examine whether visitors’ behaviours would be influenced by the seat density and types of exhibits (Table 9). The coefficient of determination R² was 0.385, indicating that exhibits and seating positions together explained 38.5% of the variance in behavioural outcomes. Multicollinearity testing showed all VIF values below 5, confirming no multicollinearity issues. Furthermore, the D-W value of approximately 2 indicated no autocorrelation in the data, with the sample data fitting the model well. Detailed analysis showed an exhibit regression coefficient of 0.288 (t = 2.278, p = 0.026 < 0.05), suggesting exhibits exert a significant positive influence on behaviour. The regression coefficient for seating was 0.146 (t = 1.196, p = 0.236 > 0.05), revealing that seating has no significant effect on behaviour.

To investigate whether the exhibition hall fosters potential social interactions, we employed DBSCAN (minimum cluster size = 2, maximum distance between clustered points = 1.5 m) to perform cluster analysis on 70 points. The results yielded 12 clusters, with the largest cluster size being 4 and all others comprising 2 points (Table 10).

Results

Heterogeneity

The most evident distinction between the two exhibitions – the permanent and temporary displays – lies in visitor flow. Further evidence is needed to determine whether admission fee has a significant impact on attendance. The VGA analysis of the temporary exhibition hall reveals that areas with higher isovist values also exhibit relatively higher integration and choice scores, suggesting a positive correlation between hall visibility and cohesion, consistent with expectations. Within the permanent exhibition, such zones predominantly occur at the junctions of exhibition modules, namely, the transition areas between ‘Designer’, ‘User’, and ‘Maker’ (Figure 7). Conversely, in the temporary exhibition, these zones are located at the exhibition’s centre, suggesting that the exhibition layout correlates with the exhibition’s thematic focus.

The relationship between VGA and dwell-social engagement

Given that prolonged visitor dwell times may cause congestion along the traffic capacity exhibition hall, seating areas are generally not positioned within highly integrated, accessible display centres (Lv et al., 2025). Spatial integration, choice options, and visual isovist do not influence visitor dwell time, suggesting that visitors focus more on exhibits and seating locations than spatial relationships. It reveals that spatial visibility exerts minimal influence on visitors’ potential social engagement, whereas spatial frequency of passage significantly impacts them.

In both exhibitions, seating is predominantly arranged in areas with low to moderate integration and choice. This approach avoids potential congestion from prolonged visitor dwell times while providing rest points during the viewing journey. However, visualisation results indicate that seating placement largely coincides with points of higher choice. In the permanent exhibition, seating density and interactive exhibits influence visitor dwell time, indirectly demonstrating that spatial choice metrics affect this behaviour. Moreover, these metrics also influence visitor social interactions, indicating that spatial accessibility shapes visitor behaviour. Regarding the integration value, the complexity of the exhibition does not affect visitor dwell time or social behaviour, which often occurs in less complex areas. Conversely, excessively high integration may reduce the likelihood of visitors lingering or engaging in conversation. Moderate or lower complexity isovists are more conducive to fostering potential social gatherings. Thus, H1 is false, and H2 is true.

Spatial affordances and social engagement

Both exhibition halls demonstrate the impact of exhibits on visitor dwell time (Tables 7 and 9), indicating that interactive area would enhance visitors’ dwell time. From Table 9, it can be seen that exhibits reveal the greatest influence on dwell time (B = 1.101, p = 0.000), followed by seating density (B = 0.551, p = 0.001), indicating that interactive exhibits and the provision of seating increase dwell time. Conversely, behaviour presents a lesser influence on dwell time, indicating that whether one views exhibits does not affect duration. During our observations, we noticed that some visitors were using their mobile phones; however, as the duration of a visitor’s stay is not directly related to whether they are viewing the exhibition, the findings indicate that there is no correlation between the staying and mobile phone use. Future research will require analysis based on a more detailed classification of behaviours.

Visitors’ dwell time is primarily determined by ‘exhibits’ and ‘seat’. Similarly, when behaviour is considered the dependent variable, the presence or absence of seating does not determine whether individuals engage with exhibits. In contrast, exhibits have a more significant impact on visitor behaviour. It should be noted that the behaviour values were B = 0.288 (towards exhibits) and B = 0.146 (towards seating). The data indicates that in areas with seating and interactive exhibits, visitors tend not to view the exhibits, which demonstrates that the phenomenon of visitors not viewing exhibits occurs more frequently in interactive zones than in static exhibition areas. In summary, interactive exhibits and seating areas would extend visitors’ dwell time, which also create opportunities for distracting behaviours.

Based on the results of DBSCAN (Table 10), although the phenomenon of coexistence has intensified, whether this can translate into actual social interaction requires more detailed interaction coding or interview studies. According to our observation, besides the exhibition halls, café, restaurant, gift shop and staircases are favoured spots for visitors to linger. The mezzanine level between the first and second floors primarily contains the museum’s studios, library and event spaces, whilst the corridor serves as display areas that link the two main galleries. A long staircase runs through the centre, inviting visitors to pause and reflect (Figure 10).

Figure 10.

Places for leisure in the museum (The authors).

Discussion

The correlative cube of VGA-DSE

In this study, we quantified the VGA metrics for two exhibition halls, primarily encompassing integration, choice, and visual isovist, and visually presented their correlations with visitors’ behaviours of dwell and social engagement. As spatial accessibility is chiefly measured by integration and choice, we define areas with high integration and high choice as high accessibility (HA), and conversely, areas with low integration and low choice as low accessibility (LA). The spatial structure of the temporary exhibition tends towards global centrality. High-visibility zones largely coincide with high-integration zones, exhibiting high accessibility and visibility (HA-HV). This layout corresponds to the exhibition’s linear narrative spatial typology, with the main storyline presented in a central position (Austin, 2020). Figure 11 demonstrates that this area features higher exhibit density, efficient spatial utilisation, and relatively concentrated content. However, this layout fails to encourage visitor dwell time and is unfavourable for fostering social gathering. Peripheral and corner zones exhibit high choice but low visibility. Seating arrangements predominantly cluster here, showing a positive correlation with interactive exhibits. However, limited sample sizes preclude establishing causal links with visitor behaviour. In contrast, permanent exhibitions display more pronounced spatial characteristics.

Figure 11.

Exhibits in the HA-HV area and seats in the hall (The authors).

The three exhibition zones within the permanent exhibition are characterised as follows: The Designer zone features low integration and choice with high visual isovist, representing a low accessibility with high visibility (LA-HV) structure; the User section shows a high integration and choice with medium isovist, embodying a high accessibility with medium visibility (HA-MV) structure; and the Maker section, like the Designer zone, is classified as an LA-HV structure. Within this exhibition, seating density and interactive exhibits significantly influence visitor dwell time. Consequently, we propose that the LA-HV structure impacts visitors’ willingness to linger. It should be noted that the largest cluster centres on the exhibition hall entrance. As an entry point, this section offers high choice, featuring an exhibit comprising over 200 everyday objects submitted by visitors from more than 25 countries (Wilson, 2016), among which are several iconic designs, such as Britain’s classic Underground Logo and IKEA shopping bags. This section vividly illustrates the mutual interaction between design and audience, with visitors naturally becoming part of the exhibition’s curatorial concept (Figure 12). From a spatial perspective, the clusters presented here suggest that medium or low complexity facilitates social engagement. The result may also stem from the exhibits’ close connection to visitors’ daily lives, effectively capturing their attention. The value of integration measures the closeness from an original space to other spaces (Teklenburg et al., 1993).

Figure 12.

The installation at the entrance (The authors).

However, visitors’ behaviour is more decided by the value of choice, the index to measure the potential accessibility of routines and related to traffic flow (Hillier et al., 1987). It means that, compared to physical spatial accessibility, visitors are more concerned with whether the pathway design facilitates smoother access to various locations. Consequently, in this case study, we have established choice as the primary metric for spatial accessibility. Where choice levels are high, the area is deemed to possess high accessibility. We thus conclude a cube to clarify the correlation between the VGA and visitors’ dwell (D) or social engagement (SE) (Figure 13).

Figure 13.

The correlative cube of VGA-DSE (The authors).

Design implications

Lefebvre (1991) advocated the interaction between space and human experience, emphasising the dynamism of space and human agency. As visitors move through a space, they exert a counterforce upon it, that is, the concept of negative space. Zhou and Zheng (2024) conceptualised the urban spatio-temporal behavioural landscape as a coupled relationship among space, time, and human activity, indicating that space, temporal rhythms and human behaviour were mutually constitutive.

In this study, we examine the mechanism between the spatial affordance and visitors’ behaviour by using the approach of space syntax. Based on the results, a ‘visibility-accessibility-dwell’ (V-A-D) supply chain is summarised, which clarifies that the visitors’ willingness for dwell is influenced by the visibility (measured by visual isovist) and accessibility (measured by integration and choice metrics) of the space, and highlights the key elements that encourage visitors to linger and interact. For example, the permanent exhibition adopted a decentralised and modular structure, with the LA-HV area encouraging visitors to linger by offering unobstructed views while minimising crowd interference. This finding provides inspiration for the exhibition’s circulation design: avoiding an excessive number of choices along the main pathways to prevent overcrowding from reducing the intensity of visitor engagement, particularly at entrances and exits. Furthermore, ‘accessibility’ also indicates the seating density and interactivity of exhibits, enhancing visitor’s engagement with the exhibition hall. By contrast, the layout of a physical space has little connection with social engagement.

The impact of seating arrangements on visitor dwell time varied between the two exhibitions. In temporary exhibitions, no correlation was found between the two. Conversely, in permanent exhibitions, seating density significantly impacts visitor dwell time. Another factor influencing behaviour through seating is that exhibits near seating are often interactive, demonstrating how micro-facilities in museums can align spatial availability with visitor needs (Achiam et al., 2014; Falk, 2016). It should be noticed that in the permanent exhibition, where the largest gathering cluster forms near the exhibition hall entrance, not merely as a necessary passageway, but because over a hundred popular everyday objects attract visitors to pause and observe. Findings reveal that in addition to physical facilities, the content of the exhibits is another factor that potentially increases visitor dwell times.

Active exploration lies at the heart of Gibson’s theory, emphasising the interaction between the environment and the perceiver, favouring the study of environment-perception relationships (Tröndle, 2014). According to the results, given that seating and exhibit design can increase visitor dwell time, interactive displays may be incorporated at key points along visitor routes, complemented by seating or spaces where visitors can linger (Tzortzi, 2017). Findings also indicate that visitor dwell time does not equate to exhibition engagement, meaning that the duration of stay shows a weak correlation with exhibition participation. It suggests that there is no significant correlation between whether visitors view the exhibition and their dwell times. Although visitors tend to linger longer in interactive zones, the probability of not viewing the exhibition is actually higher.

Falk’s theory on motivations for museum visits further corroborates this point, identifying a distinct category of visitors termed ‘rechargers’-individuals who attend museums not primarily to view exhibits, but to rest and unwind (Falk, 2011). The other four visitor categories comprise the explorer (seeking new knowledge), the facilitator (engaging socially), the experience seeker (collecting and documenting experiences), and the professions (engaging in active critique). Thus, we can consider that visitors who do not engage with the exhibits can also be regarded as participating in the museum experience. However, our observations of the permanent exhibition reveal that the entrance area, featuring display walls and an overview of design chronology, tends to attract concentrated visitor flows. According to clustering results, high co-presence indices do not necessarily indicate high social potential; rather, they may impede the space’s capacity for movement. While technological solutions can be employed to enhance exhibition interactivity and visitor engagement, strategies for visitor flow management to improve spatial mobility should also be considered to prevent potential crowding resulting from extended dwell times.

Limitations and future research

This paper focuses on how spatial configurations within museums influence visitor dwell times and social engagement, exploring a research framework that integrates space syntax with behavioural observation. The limitations of this study are as follows: (1) Single-case extrapolation. The framework for spatial affordances influencing visitor behaviour was derived from a case study of the London Design Museum. Future research should validate these findings through comparative analyses across multiple venues and regional case studies. (2) The paper focuses on examining the relevance and trends of spatial factors within museums using the method of space syntax. Data collection occurred on 1 day of weekdays with a limited data volume. Moreover, there is a significant difference in the number of samples between the two exhibition halls. The findings may not support causal inferences, potentially introducing seasonal variations and data bias. (3) Limited behavioural categorisation. This paper employs cluster analysis and defines sociological social distance to identify factors potentially influencing visitors’ social interactions within venues. This methodology quantifies spatial provision and visitor behaviour, providing a basis for exhibition layout. However, it is impossible to accurately determine the distribution of their attention through non-interventional observation. More demographic information needs to be collected for researching visitor behaviour in details.

Future research should conduct categorised behavioural studies, combined with in-depth interviews with different stakeholders like architects and curators, to assess the hierarchical analysis between individual backgrounds, venue layout, and cultural factors. Data volume will be supplemented to strengthen the causal relationship. More comprehensive analysis of the of the museum’s spatial and urban surroundings will be addressed in the future research agenda.

Conclusion

This paper analyses the case of tDM using a mix method. Space syntax is applied to quantify the relationship between facilities and visitors’ behaviours in two exhibitions. Combing Depthmap and QGIS, the value of integration, choice and isovist is visualised, in which the correlation among them can be directly seen. Concurrently, it enables spatial assessment of potential social interactions within both exhibition halls and identifies areas most likely to facilitate such engagement. SPSS statistical analysis is employed to examine the correlation between spatial VGA, seating arrangements and exhibit types, alongside the relationship between facilities and visitors’ dwell times and social behaviours. Findings indicate that exhibits exert a greater influence on visitor dwell time and social interaction than the spatial environment.

Our observations indicate that cafés and restaurants are also areas where people tend to linger. Furthermore, seating and interactive exhibits enhance the duration of visitors’ stays within museums, whilst whether or not they view the exhibitions has no significant impact on their stay. However, whether this ‘staying’ signifies rest or learning requires more detailed investigation. Analysing the significance of museums to audiences necessitates focusing on spatial typologies, visitor types and behavioural subcategories, which also provides a method for researching how museums might offer diverse visiting programmes.