A Cognition–Affect–Behavior Framework for Assessing Street Space Quality in Historic Cultural Districts and Its Impact on Tourist Experience

Abstract

Existing research predominantly focuses on the preservation or renewal models of the physical forms of historic cultural districts, with limited exploration of their roles in stimulating tourists’ cognitive, affective resonance, and behavioral interactions. This study addresses historic cultural districts by evaluating the space quality and its impact on tourist experiences through the “cognition-affect-behavior” framework, integrating GIS, street view semantic segmentation, VR eye-tracking, and web crawling technologies. The findings reveal significant multidimensional differences in how space quality influences tourist experiences: the impact intensities of functional diversity, sky visibility, road network accessibility, green visibility, interface openness, and public facility convenience decrease sequentially, with path coefficients of 0.261, 0.206, 0.205, 0.204, 0.201, and 0.155, respectively. Additionally, space quality exerts an indirect effect on tourist experiences through the mediating roles of cognitive, affective, and behavioral dimensions, with a path coefficient of 0.143. This research provides theoretical support and practical insights for empowering cultural heritage space governance with digital technologies in the context of cultural and tourism integration.

1. Introduction

As the physical carrier and space representation of urban cultural memory, historic cultural districts have become a critical topic in the synergistic development of preservation, renewal, and adaptive reuse within the process of new urbanization [1,2]. With the deepening integration of culture and tourism, tourist experiences have gradually extended beyond the traditional aesthetic appreciation of landscapes, shifting toward a deeper perception of space quality and the spirit of place [3,4]. However, existing studies have primarily focused on the preservation of physical forms [5] or the evaluation of built environment values [6], with insufficient attention paid to the interaction between tourists’ psychological and behavioral mechanisms and space quality. Empirical analyses based on multimodal intelligent analytical methods are particularly lacking.

The “Cognition-Affect-Behavior” theory, originating in the field of psychology, can be traced back to Ellis’s ABC (Activating events–Beliefs–Consequences) affective theory proposed in the 1950s [7] and Beck’s cognitive theory in the 1960s [8]. Later, in 1997, Baloglu and Brinberg introduced the “Cognition-Affect” model [9] to explain how tourists’ cognitive evaluations and affective responses to destination images collectively shape their overall perceptions. In recent years, this theory has been increasingly applied in the field of architecture to reveal how environmental stimuli influence individual behavioral decisions through cognitive processing and affective feedback.

In the context of tourist experiences in historic districts, cognition serves as the foundation, referring to tourists’ perception and recognition of both tangible and intangible elements of the district. Affect acts as the critical link between cognition and behavior, while behavior represents the ultimate manifestation of tourist experience. The “Cognition-Affect-Behavior” theory provides a comprehensive multidimensional framework for understanding the complexity of space experiences, uncovering the dynamic relationships between space quality and tourists’ cognition, affect, and behavior, thereby offering an essential conceptual foundation for understanding the mechanisms through which space quality impacts tourist experiences.

Cognitive objects can be categorized into four types: architectural landscapes, artistic landscapes, living landscapes, and public landscapes [10]. Architectural landscapes, as the material carriers of the district, convey historical information and cultural connotations through elements such as form, material, and scale [11,12]. Artistic landscapes encapsulate aesthetic value and artistic expression, serving as focal points for tourists’ visual and aesthetic experiences [13,14]. Living landscapes reflect the vibrancy and locality of urban life, embodying the everyday experiences of residents [15,16]. Public landscapes provide space venues for social interactions and cultural activities, fostering interactions between tourists and residents, as well as among tourists themselves [17,18]. The cognitive process lies at the core of tourist experiences, forming the basis for affective responses.

Tourist affects can be classified into positive and negative affects. Positive affects, such as pleasure, excitement, satisfaction, and nostalgia, form the foundation for cultural attachment and place identity [19,20]. Negative affects, such as disappointment, confusion, and frustration, often stem from discrepancies between expectations and reality or environmental discomfort [21,22]. The relationship between cognition and affect is context-dependent; the same landscape element may elicit different affective responses depending on tourists’ cultural backgrounds, travel motivations, and prior knowledge [23].

Tourist behaviors can be divided into self-sufficient behaviors and constructive behaviors [24]. Self-sufficient behaviors are passive adaptive actions undertaken by tourists to meet individual needs, such as sightseeing, photography, and resting [25]. Constructive behaviors involve active participation, creation, and dissemination based on in-depth experiences, such as shopping, cultural learning, and word-of-mouth sharing [20]. Positive affects significantly drive constructive behaviors and further enhance these tendencies by strengthening tourists’ sense of place attachment and immersion [26].

This study focuses on historic cultural districts, employing the “Cognition-Affect-Behavior” theoretical framework to develop a multidimensional space quality evaluation system encompassing street traffic, street functions, street facilities, and street environments. By integrating web text analysis and VR eye-tracking experiments, the study analyzes tourist experiences and reveals the impact of space quality on these experiences. The objectives of this study are to: (i) Uncover the formation mechanism of tourist experiences through the dynamic chain of “cognitive triggering—affective feedback—behavioral response”; (ii) integrate heterogeneous data sources and interdisciplinary technical methods to achieve complementary validation of subjective and objective evaluations; and iii) apply Nanchang’s Wanshou Palace as a case study to provide a replicable analytical framework for improving space quality and optimizing experiences in similar historic cultural districts.

2. Related Work

2.1. Evaluation of Space Quality in Historic Cultural Districts

Space quality represents a comprehensive assessment of the overall characteristics of the urban physical environment, aiming to address the multidimensional needs of human interaction with space [27]. At the street level, the evaluation of space quality focuses on the space attributes of streetscapes, with methodologies evolving from small-scale empirical studies to large-scale quantitative analyses alongside technological advancements. Early studies predominantly relied on traditional methods such as field surveys, direct observation, and interviews. While these approaches allowed for an in-depth exploration of space characteristics, they were highly time-consuming and labor-intensive, thereby limiting their applicability to small-scale areas [28,29]. In recent years, the integration of street view imagery with deep learning technologies has facilitated the rapid development of large-scale quantitative research [30,31]. By leveraging street view imagery data sourced from platforms such as Google Maps, Baidu Maps, and Tencent Maps, and applying deep learning techniques for batch data processing, researchers have been able to efficiently conduct large-scale evaluations of street space quality. Furthermore, the adoption of biosensor technologies (e.g., eye-tracking devices and electroencephalograms) has provided novel methodological pathways for investigating space perception by monitoring physiological responses [32,33].

The evaluation framework for space quality encompasses a range of indicators, which can be categorized into single-element and multi-element metrics based on their content [34]. Single-element metrics focus on individual space components, such as greenery or sky visibility, with common indicators including the green visibility and sky visibility. In contrast, multi-element metrics address the interplay of multiple space features, such as buildings, roads, and walls, with representative indicators including interface openness. From the perspective of evaluation, space quality can be further divided into objective quality and subjective needs [35]. Objective quality reflects space morphological characteristics derived from measurable elements, which in turn evoke distinct visual experiences. Subjective needs, on the other hand, exhibit variability due to individual differences and are challenging to quantify directly, often encompassing psychological dimensions such as comfort, safety, and aesthetic pleasure. Additionally, space quality indicators can be classified into four dimensions based on street content: transportation, functionality, facilities, and environment [36]. The transportation dimension evaluates road network accessibility, the functionality dimension emphasizes functional diversity, the facilities dimension focuses on convenience, and the environmental dimension assesses landscape diversity. By employing a multidimensional and hierarchical system of indicators, research on space quality can comprehensively capture the complexity and diversity of urban streetscapes.

The measurement of space quality indicators reflects a multidimensional approach, shaped by differences in cognitive processes and the diversity of data sources. In existing studies, certain indicators—such as the green visibility, sky visibility, interface openness, and public facility convenience—are well-defined and can be quantified by assessing the visual proportion or space distribution of environmental elements, thereby providing a foundation for objective evaluations of space quality [37,38]. For instance, green visibility is calculated as the proportion of green vegetation within the visual field, reflecting the visual comfort and ecological quality of the environment. Similarly, sky visibility measures the proportion of visible sky, serving as an indicator of space openness and transparency. However, for indicators such as road network accessibility and functional diversity, the limitations of street view imagery often necessitate the integration of additional data sources for comprehensive analysis [39,40]. Road network accessibility is typically quantified through line density analysis of path data combined with GIS techniques, enabling the evaluation of transportation connectivity and space accessibility. Functional diversity, on the other hand, is assessed through a kernel density analysis of Points of Interest (POI) data, capturing the degree of functional mixing and the space distribution of commercial, cultural, and recreational activities. The introduction of machine learning models has further enhanced the precision of these measurements, enabling more robust and scalable evaluations.

2.2. Research on Tourist Experience Effects Based on Computer Vision Technology

Research on tourist experiences has progressively developed into a multidimensional theoretical framework. Early studies predominantly focused on the subjectivity and interactivity of experiences, conceptualizing tourist experiences as dynamic interactions between tourists and destinations [41]. With the advancement of research, scholars have expanded the theoretical scope from perspectives such as satisfaction, benefits, and the essence of experiences, proposing that tourist experiences are the pleasurable sensations and psychological fulfillment derived from personal engagement during the tourism process [42]. Simultaneously, the content of tourist experience research exhibits significant space and temporal differentiation. From a space perspective, different types of tourism destinations (e.g., rural areas [43], commercial complexes [44], and historical districts [45]) distinctly shape tourist experiences. From a temporal perspective, specific time periods (e.g., nighttime [46] and festive occasions [47]) significantly influence tourists’ perceptions and behavioral patterns. These space and temporal variations offer diverse research perspectives and practical pathways for the study of tourist experiences.

The application of technologies such as computer vision, affective computing, and behavioral tracking analysis in tourist experience research has become increasingly widespread, enabling the multimodal quantification of urban spaces and tourist experience analysis. Deep learning techniques (e.g., CNNs and Transformer architectures) [48,49,50], through training on large-scale image datasets, accurately extract characteristic patterns of environmental elements such as building facades, greenery, and the sky, providing robust technical support for the quantitative analysis of space quality. Semantic segmentation techniques [51,52] enable the precise classification of image content, allowing for a detailed analysis of functional zones and landscape elements within spaces, thereby uncovering the specific impacts of space features on tourist perceptions. Meanwhile, eye-tracking technologies [53,54,55], utilizing VR/AR devices to record the dynamic changes in tourists’ visual attention, assist researchers in understanding tourists’ visual preferences and their associations with cognitive processes. Sentiment analysis of tourist reviews [56,57], leveraging natural language processing techniques, extracts affective tendencies and experiential feedback from user-generated textual data, offering a critical basis for the quantitative evaluation of tourist experiences.

3. Method

3.1. Research Area

This study takes the Wanshou Palace Historic and Cultural District in Nanchang City, Jiangxi Province, as a case study. As one of the first provincially designated historic and cultural districts in Jiangxi, it is enclosed by Zhongshan Road to the north, Cuihua Street to the east, and Chuanshan Road to the southwest (Figure 1). The district comprises nine streets and alleys, along with three sunken courtyards, forming an internal space structure characterized by “four main streets, five alleys, six horizontal and three vertical axes.” The entire district is divided into three functional zones based on its two main streets, Jingyang Street and Hetong Alley: Zone A (Zhongshan Road–Jingyang Street area, featuring multi-story buildings), Zone B (Chuanshan Road core commercial area, characterized by low-rise buildings), and Zone C (Wanshou Palace Museum area, dominated by palace-style architecture). As shown in Figure 1, the fifteen scenes represent typical scenarios from three functional zones, with varying street morphologies that effectively reflect the overall characteristics of the zones. Specifically, scenes 5, 6, and 10 belong to Zone A; scenes 1–4, 7–9, and 14–15 belong to Zone B; and scenes 11–13 belong to Zone C. So, the fifteen representative scenes were selected for VR eye-tracking experiments to comprehensively capture the space characteristics of the district. The streetscape images of these scenes will be utilized in subsequent VR eye-tracking experiments to analyze tourists’ visual attention distributions and cognitive preference variations.

3.2. Research Framework

This study establishes a comprehensive analytical framework for the space quality and tourist experience of historic and cultural districts from the perspective of “cognition-affect-behavior” (Figure 2). First, a space quality evaluation model is developed by integrating multi-source data, encompassing four dimensions: transportation, functionality, facilities, and environment. Second, tourist experiences are analyzed through a combination of eye-tracking technology and content analysis. Finally, using survey questionnaire data, a structural equation model (SEM) is employed to validate the transmission mechanism of the space quality’s impact on tourist experiences. Ultimately, a chain model of tourist experience—“cognitive trigger–affective feedback–behavioral response”—is constructed.

3.3. Construction and Evaluation Methods of the Space Quality Indicator System

Based on the theoretical framework of space quality research both domestically and internationally, this study establishes a multi-level space quality evaluation indicator system, comprising four primary indicators (street transportation, street functionality, street facilities, and street environment) and six secondary indicators (road network accessibility, functional diversity, public facility convenience, green visibility, sky visibility, and interface openness). Road network accessibility measures the convenience of street transportation, where high accessibility enhances pedestrian reachability and activity efficiency. Functional diversity, achieved through the rational allocation of various functions to meet diverse needs, serves as a core dimension for evaluating street vitality and usage efficiency. Public facility convenience assesses the overall service level of transportation, rest areas, sanitation, and landscape facilities. The green visibility and sky visibility reflect street greening levels and visual transparency, respectively, serving as key indicators of environmental comfort. Interface openness represents the space openness of streets, directly influencing pedestrian activity experiences.

The methods for quantifying space quality indicators are as follows: road network accessibility is determined through a line density analysis of path data, while functional diversity is evaluated using a kernel density analysis of POI data. Public facility convenience, green visibility, sky visibility, and interface openness are calculated based on the semantic segmentation results of streetscape images. Specifically, path data and POI data are collected through Python 3.8 web scraping and spatially analyzed using the ArcGIS Pro 3.0 platform. A total of 203 valid streetscape image samples were collected through field surveys and clarity screening (September 2024, daytime), which were processed using deep learning algorithms for semantic segmentation. The proportions of various elements were calculated, and the segmentation results were spatialized and visualized by integrating geographic coordinate information. The calculation formulas for each indicator are as follows:

Road Network Accessibility (RNA):

$$RNA={\displaystyle \frac{L}{A}}$$

(1)

In the above formula, $L$ represents the total length of road centerlines within the space unit (unit: meters), while $A$ denotes the total land area of the space unit (unit: square meters).

Functional Diversity (FD):

$${FD}_{\left(t\right)}={\sum }_{i=1}^n{\displaystyle \frac{1}{h^2}}\mathrm{K}\left({\displaystyle \frac{d_i}{h}}\right)$$

(2)

For POI point features, kernel density analysis is employed to calculate the density of point features within the vicinity of each output raster cell. In the above formula, ${FD}_{\left(t\right)}$ represents the estimated density of functional diversity, $t$ denotes the central coordinates of the output raster cell, $n$ is the number of point features (unit: count) located within the search area centered at $t$ with a radius of $h$, $K$ is the kernel function, and $h$ refers to the bandwidth (unit: meters).

Public Facility Convenience (PFC):

$${PFC}_n={\displaystyle \frac{P_n}{A_n}}={\displaystyle \frac{{\sum }_{i=1}^{\iota }{p}_i}{{\sum }_{i=1}^i{a}_i}}$$

(3)

In the above formula, $P_n$ represents the total number of pixels occupied by service facilities in the n-th streetscape image, which is the sum of the pixels corresponding to the i-th facility areas within the image (unit: pixels). $A_n$ denotes the total number of pixels in the streetscape image, representing the sum of all area pixels within the image (unit: pixels).

Green Visibility (GV):

$${GV}_n={\displaystyle \frac{G_n}{A_n}}={\displaystyle \frac{{\sum }_{i=1}^i{g}_i}{{\sum }_{i=1}^i{a}_i}}$$

(4)

In the above formula, $G_n$ represents the total number of pixels occupied by trees and vegetation in the n-th streetscape image, which is the sum of the pixels corresponding to the i-th vegetation areas within the image (unit: pixels); $A_n$ denotes the total number of pixels representing all area features in the streetscape image (unit: pixels).

Sky Visibility (SV):

$${SV}_n={\displaystyle \frac{S_n}{A_n}}={\displaystyle \frac{{\sum }_{i=1}^i{s}_i}{{\sum }_{i=1}^i{a}_i}}$$

(5)

In the above formula, $S_n$ represents the total number of pixels occupied by the sky in the n-th streetscape image, which is the sum of the pixels corresponding to the i-th sky areas within the image (unit: pixels). $A_n$ denotes the total number of pixels representing all area features in the streetscape image (unit: pixels).

Interface Openness (IO):

$${IO}_n=1-{\displaystyle \frac{{\sum }_{i=1}^i{m}_i}{{\sum }_{i=1}^i{a}_i}}$$

(6)

In the above formula, ${IO}_n$ represents the interface openness of the n-th streetscape image, $m_i$ is the number of pixels in the i-th enclosure area within the image (unit: pixels), $a_i$ is the total number of pixels in the i-th area of the streetscape image (unit: pixels), and ${\displaystyle \frac{{\sum }_{i=1}^i{m}_i}{{\sum }_{i=1}^i{a}_i}}$ represents the degree of interface openness.

3.4. Tourist Experience Effect

From the perspective of the “Cognition-Affect-Behavior” framework, this study builds a multidimensional evaluation index system for tourist experiences, drawing on the research by Qin Zhaoxiang et al. on tourists’ visual perception and affective experiences in historical and cultural districts. The cognitive dimension encompasses tourists’ visual perception preferences for different landscape types (architectural landscapes, artistic landscapes, living landscapes, and public landscapes); the affective dimension includes both positive and negative affects; and the behavioral dimension is divided into self-sufficient behaviors and constructive behaviors.

To quantify the multidimensional characteristics of tourist experiences, the study integrates VR eye-tracking experiments with online review data analysis to systematically analyze tourists’ cognitive, affective, and behavioral feedback. First, a VR eye-tracking experiment (Figure 3) was conducted, recruiting 30 volunteers with academic backgrounds in architecture and urban–rural planning. Using VR eye-tracking devices, participants observed streetscape images of 15 typical scenes (60 s per scene), with synchronized recordings of fixation heatmaps and saccade paths. The experimental parameters were set as follows: a fixed duration of 2 s for saccade paths, a fixed number of 10 gaze points, a heatmap hotspot duration of 5 s, and a hotspot turning red after 800 milliseconds. These parameters aimed to capture the distribution of tourists’ visually sensitive areas for architectural, artistic, living, and public landscapes, enabling a comparative analysis of the cognitive impacts of different landscape types. After viewing, participants rated the 15 typical scenes on three aspects—“the extent to which this scene enhances your understanding of the Wanshou Palace district, your favorability toward it, and your willingness to visit”—on a scale of 1 to 5. These ratings provided data support for exploring the impact mechanisms of space quality on tourist experiences. Simultaneously, 1720 review comments from platforms such as Dianping and Ctrip (spanning February 2024 to February 2025) were scraped using Python. After segmentation with Jieba and the filtering of stop words, the TF-IDF algorithm was applied to extract high-frequency feature words, quantifying the tourists’ affective feedback and behavioral tendencies, thereby enabling a fine-grained analysis of tourist experiences.

To eliminate the interference of visual area size on fixation duration and objectively evaluate the visual attractiveness of environmental elements, this study introduces Attention Percentage (AP) as a quantitative indicator. The calculation formula for Attention Percentage is as follows:

$$AP={\displaystyle \frac{Di}{{\sum }_n^{j=1}Di}}$$

(7)

$Di$ represents the cumulative fixation duration for the i-th element, while $\sum Di$ denotes the total cumulative fixation duration for all n elements across 15 scenes. To validate the effectiveness of this metric, the study randomly selected five volunteers and conducted a frame-by-frame analysis of their VR tour videos across the 15 scenes, recording their fixation sequences and cumulative fixation durations for different elements. By calculating the proportion of each element’s cumulative fixation duration relative to the total fixation duration, the visual attractiveness of each element was quantitatively assessed.

To systematically explore the mechanisms through which space quality influences tourist experiences, the study employed structural equation modeling (SEM) and mediation effect analysis using Amos 26.0 and the SPSS plugin Process 3.1. Based on six core indicators of space quality (road network accessibility, functional diversity, public facility convenience, green visibility, sky visibility, and interface openness), a theoretical model of “space quality → cognition → affect → behavior” was constructed. The model’s fit was evaluated using goodness-of-fit indices (e.g., CFI, TLI, and RMSEA) to ensure its theoretical validity and statistical robustness. Finally, potential mediation and moderation effects were further examined to uncover the multi-level pathways through which space quality impacts tourist experiences.

4. Results and Discussion

4.1. Analysis of Space Quality Evaluation

The space quality of the study area exhibits significant variation across the three functional zones (Figure 4). Zone A, encompassing the Zhongshan Road–Jingyang Street area (Scenes 5, 6, and 10), serves as a transportation hub adjacent to subway stations, characterized by a well-developed signage system, abundant public landscapes, and high public facility convenience (Figure 5a). This area is dominated by high-rise buildings, with architectural landscapes accounting for 26.1%. However, the scarcity of ground-level commercial spaces and limited presence of everyday landscapes, such as shop signs, result in low functional diversity (Figure 5a). The wide street cross-sections and sparse greenery contribute to high sky visibility (average 28.3%) and interface openness (average 72.5%), creating a spacious and transparent space quality (Figure 5b).

Zone B, the Chuanshan Road core commercial area (Scenes 1–4, 7–9, and 14–15), functions as the central business district with a high-density road network (road density of 5.2 km/km²). This area features numerous public landscapes, such as signage, and exhibits high public facility convenience (Figure 5c). The combination of artistic decorations and lifestyle-oriented commercial layouts (68% commercial spaces, 24.2% architectural landscapes) enriches the functional diversity of this zone (Figure 5d). The inner alleyway spaces are characterized by low sky visibility and minimal greenery (0–5%, Figure 5e); however, the design of three-dimensional space structures and plaza nodes enhances sky visibility (Figure 5f). Additionally, the historical greenery along Chuanshan Road improves the green visibility (Figure 5g).

Zone C, the Wanshou Palace Museum area (Scenes 11–13), serves as a cultural exhibition hub with abundant architectural landscape resources. Historical buildings such as Wanshou Palace and the opera stage, along with artistic landscapes like sculptures and murals, further reinforce the cultural atmosphere. Morphologically, this zone maintains high sky visibility (average 26.5%) and interface openness (average 78.4%, Figure 5h). The internal greenery and public landscape design within Wanshou Palace create localized peaks in the green visibility (12%, Figure 5i). However, the single-sided commercial layout on Cuihua Street (Figure 5j) limits the diversity of everyday functions in the area.

Based on the functional and morphological characteristics of the spaces, the study identifies three typical space types: gateway spaces (e.g., Scene 3: Chuanshan Road, Scene 6: Zhongshan Road), plaza spaces (e.g., Scene 8: Zhongshen Well, Scene 12: Dongfu Well), and alleyway spaces (e.g., Scene 4: Qiaobu Street, Scene 11: Cuihua Street). (1) Gateway spaces prioritize transportation convenience and well-developed facilities, although accessibility is influenced by the nature of the roads. These spaces typically exhibit high public facility convenience and a significant proportion of public landscapes. (2) Plaza spaces emphasize space openness and visual appeal. For instance, Scenes 8 and 12 both demonstrate high openness due to their three-dimensional architectural layouts. However, the degree of enclosure affects their attractiveness to crowds; Scene 8 features an enclosed layout with an inward space orientation, while Scene 12 adopts a semi-enclosed layout that is more outward facing, attracting larger crowds. (3) Alleyway spaces create functional diversity through the integration of commercial and artistic landscapes, influenced by varying everyday landscape patterns. For example, double-sided commercial layouts (e.g., Scene 4) attract customers with a strong commercial atmosphere, whereas single-sided layouts (e.g., Scene 11) rely on artistic landscapes such as murals to draw visitors. The differentiated characteristics of these three space types collectively contribute to the diversity of space functions and forms, providing valuable references for the optimization of different scenes.

4.2. Analysis of Tourist Experience Results

4.2.1. Analysis of the Cognitive Dimension

Through the VR-based tourist experience experiment, the impact of different landscape types on tourists’ cognitive perceptions across 15 typical scenes was analyzed (Figure 6): (1) architectural landscapes primarily exhibit visual appeal through historical and cultural elements. Iconic structures, such as the Tie Zhu Wanshou Palace and Guangrun Gate, attract a prolonged fixation through color contrasts, textual information, and historical imprints. Features like traditional Chinese characters and weathered stone walls further extend gaze duration. In contrast, ordinary historical buildings garner moderate attention through wooden structures and signage information. (2) Cognitive responses to artistic landscapes display a polarized pattern. Traditional murals and sculptures elicit brief attention but often lead to visual fatigue, causing focus to shift to narrative elements such as human activities. Conversely, modern cultural walls, with their distinctive features and interactive designs, sustain prolonged engagement, particularly among younger audiences. (3) The cognitive perception of everyday landscapes is closely related to space positioning and information recognition. Projecting shop signs, due to their prominent placement and decorative features, become primary visual focal points, while ground-level advertisements and market stalls extend gaze duration through close-range information. (4) Public landscapes are influenced by space layout. Greenery and corridors form localized visual focal points; however, their singular forms often fail to sustain prolonged attention, with gaze frequently shifting toward surrounding buildings or the sky. Overall, the visual appeal of landscape elements is positively correlated with their formal uniqueness, informational richness, and space accessibility. Conversely, overly simplistic or excessively complex designs may induce visual fatigue, thereby diminishing the effectiveness of landscape cognition.

The attentional distribution across different types of landscapes exhibits significant space variation (

Table 1. Proportion of fixation duration.

Environmental Elements		Percentage of Attention
Architectural Landscape	Architecture	38.4%	41.6%
	Architectural Details	3.2%
Artistic Landscape	Decoration	2.3%	7.6%
	Mural	2.8%
	Sculpture	2.5%
Living Landscape	Shop	22.7%	29.9%
	Shop Sign	7.2%
Public Landscape	Crowd	0.8%	8.9%
	Greenery	6.5%
	Traffic Sign	1.6%
Non-landscape Elements	Sky	1.5%	12.0%
	Road	5.3%
	Others	5.2%

) and temporal characteristics (Figure 7). Architectural landscapes garnered the highest level of attention (41.6%), followed by everyday landscapes (29.9%), while public and artistic landscapes demonstrated relatively weaker appeal (accounting for 8.9% and 7.6%, respectively). This pattern reflects a progressive cognitive process in tourists’ perception of landscape elements, transitioning from the overall structure to specific details, and from functional aspects to intricate features: initial attention is drawn to the overall architectural form, followed by textual information, and ultimately focused on decorative details. This cognitive model provides critical insights into the landscape design of historic districts, suggesting that while preserving the overall aesthetic integrity, emphasis should be placed on enhancing the detailed design of key nodes to prolong tourists’ visual engagement.

4.2.2. Analysis of the Affective Dimension

An analysis of online content reveals that tourists’ affective responses to the Wanshou Palace Historic and Cultural District are characterized by a mixture of positive and negative sentiments, with a predominance of positive affects. Notable differences are observed between cultural and functional experiences (

Table 2. High-frequency words in the comments.

Cognitive Entry	TF-IDF	Affective Entry	TF-IDF	Behavioral Entry	TF-IDF
Wanshou Palace	0.08636	Bustling	0.04370	Walking	0.06519
Historic Building	0.04549	Crowded	0.04147	Visiting Landmark	0.06015
Commercial Street	0.04112	Good	0.03724	Visiting a Night Market	0.05318
Delicacies	0.03894	Delicious	0.03191	Snacking	0.04573
Social Media Popular Shop	0.02637	Unique	0.03103	Taking Photos	0.03647
Subway	0.02123	Worthwhile	0.02298	Drinking Tea	0.03455
Night View	0.02098	Good-looking	0.01955	Show	0.02715
Taoism	0.01470	Convenient	0.01656	Singing	0.02693
Parking Lot	0.01256	Expensive	0.01487	Buying Cultural	0.01185
Temple	0.00833	Commercialized	0.01130	Sitting Quietly	0.00872
Zhongshan Road	0.00723	Enjoyable	0.00842	Listening to a Guided Tour	0.00691
Xu Xun	0.00617	Average	0.00826	Burning Incense and Praying	0.00642

). In terms of cultural experiences, “Wanshou Palace,” as a core cultural symbol, is closely associated with keywords such as “Taoism” and “temple,” collectively fostering a strong sense of cultural identity and eliciting positive affects such as “worthwhile” and “pleasant.” Regarding functional experiences, everyday landscapes such as “historic buildings,” “commercial streets,” and “local cuisine” are frequently linked with high-frequency terms like “lively,” “delicious,” and “unique,” reflecting tourists’ favorable evaluations of the district’s atmosphere and functionality. However, the negative impact of over-commercialization cannot be overlooked. High-frequency terms such as “expensive” and “crowded” highlight the detrimental effects of excessive commercialization and high visitor density on the quality of the experience. Although transportation facilities such as the “subway” and “parking lots” enhance accessibility, the additional foot traffic generated by the Zhongshan Road commercial area exacerbates the overcrowding within the district.

4.2.3. Analysis of the Behavior Dimension

Behavioral analysis based on high-frequency terms extracted from reviews (Table 2) reveals two primary patterns of tourist behavior. Self-sufficient behaviors, such as “strolling,” “sitting quietly,” and “taking photos at check-in spots,” reflect tourists’ spontaneous acceptance of the district’s environment. Constructive behaviors, including “eating snacks,” “purchasing cultural and creative products,” and “burning incense for blessings,” demonstrate active participation stimulated by environmental features. Together, these two behavioral patterns constitute the core of tourists’ place-based experiences.

The formation of these behavioral patterns follows a progressive “cognition-affect-behavior” mechanism. Tourists initially establish a preliminary understanding of the district through visual perception and historical cognition, which subsequently triggers affective responses. Positive affects, such as “lively” and “unique,” encourage deeper engagement behaviors, whereas negative experiences, such as “crowded” and “expensive,” may inhibit consumption behaviors. Notably, behavioral tendencies exhibit significant scene dependency: in cultural nodes like Wanshou Palace, tourists are more inclined toward cultural experience activities such as “watching performances” and “burning incense for blessings,” while in commercial areas, behaviors such as “eating snacks” and “purchasing cultural products” are more prevalent. This highlights the close relationship between tourist behavior and the functional attributes of specific locations, offering valuable insights for optimizing the functional design of the district.

4.3. Space Quality Factors Influencing Tourist Experience

Statistical tests of the structural equation modeling (SEM) results indicate that the reliability and validity of all latent variables meet statistical requirements (

Table 3. Statistical metrics of structural equation model fit indices.

AdaptationIndex	CMIN/DF	GFI	AGFI	NFI	RFI	IFI	TLI	CFI	RMSEA
Frontage Standard	<3	>0.8	>0.8	>0.9	>0.9	>0.9	>0.9	>0.9	<0.08
Numerical Value	1.375	0.994	0.982	0.993	0.987	0.998	0.996	0.998	0.029
Matching Degree	Excellent

). Specifically, Cronbach’s α coefficients exceed 0.7, KMO values surpass 0.8, and the factor loadings of all measurement variables are above 0.7, demonstrating strong internal consistency and external validity without significant multicollinearity issues. Key fit indices, including CMIN/DF, GFI, AGFI, NFI, RFI, IFI, TLI, CFI, and RMSEA, all outperform their respective threshold criteria, indicating that the SEM exhibits a robust model fit and provides a reliable foundation for subsequent data analysis and theoretical validation.

The SEM results (Figure 8) reveal a complex multidimensional relationship between space quality and tourist experience. First, the influence of space quality components on tourist experience varies significantly: functional diversity exerts the strongest impact (path coefficient = 0.261), followed by sky visibility (0.206), road network accessibility (0.205), green visibility (0.204), interface openness (0.201), and public facility convenience (0.155), which have relatively weaker effects. This hierarchical influence reflects tourists’ core demands in historic districts: a desire for both diverse functional experiences and an enhanced sense of space quality. Second, among the three dimensions of tourist experience, the affective dimension contributes the most (0.315), followed by the cognitive dimension (0.270), with the behavioral dimension being relatively weaker (0.187). This indicates that tourists prioritize an affective resonance derived from the historical and cultural atmosphere over purely physical space experiences. Notably, although the direct impact of space quality on tourist experience is relatively modest (0.143), its overall influence remains significant through the mediating effects of the cognitive, affective, and behavioral dimensions. This mechanism underscores the unique role of space quality in historic districts: as a cultural carrier, its value lies not only in the physical space but also in its ability to evoke cultural identity and affective resonance through space design. Therefore, the optimization of historic district spaces should prioritize enhancing functional diversity while improving environmental qualities such as sky visibility and green visibility to strengthen affective experiences, ultimately achieving a synergistic enhancement of space quality and cultural value.

5. Conclusions and Outlook

This study, grounded in the “cognition-affect-behavior” theoretical framework and convileveraging multi-source data integration, systematically evaluates the space quality of the Wanshou Palace Historic and Cultural District and its impact on tourist experience. The findings are as follows: (1) Space Quality: The district demonstrates notable strengths in public facility convenience, sky visibility, and interface openness, while the green visibility requires improvement. It is recommended to optimize the greening layout through diverse strategies such as vertical green walls, landscaped flowerbeds, and rooftop gardens. Space differentiation analysis reveals that the southern section outperforms the northern section, with the latter exhibiting deficiencies in road network accessibility and functional diversity. Improvements could be achieved by optimizing vertical space layouts, adding escalators, and implementing functional repurposing strategies. (2) Tourist Experience: While the overall tourist experience in the district is positive, there remains room for improvement. Recommendations include strengthening artistic and public landscapes by incorporating art installations and optimizing rest facilities; avoiding over-commercialization to preserve the cultural ambiance; and moderately opening sections of the Wanshou Palace Museum to enrich visitors’ behavioral experiences. (3) Impact of Space Quality on Tourist Experience: The influence of space quality components on tourist experience varies, with functional diversity, sky visibility, road network accessibility, green visibility, interface openness, and public facility convenience exerting progressively diminishing impacts. Space quality primarily affects the tourist experience indirectly through the mediating effects of cognitive, affective, and behavioral dimensions. It is hypothesized that tourists experience a deeper emotional and intellectual connection to the historical and cultural elements of the district, suggesting that these intangible factors exert a direct and significant influence on visitor experiences, beyond the physical quality of the space. This finding underscores the scientific value and novelty of the study, offering a new perspective for understanding the complexity of visitor experiences.

Although the space characteristics of historic districts vary, this study constructs a multidimensional visual intelligence analysis framework to provide a novel approach for evaluating space quality. By integrating GIS, semantic segmentation, and VR technology with online review data mining, the research achieves a multidimensional analysis of space quality, enhancing the comprehensiveness and scientific rigor of the evaluation. The multi-source fusion analysis, based on trajectory data, POI data, and street view imagery, further enriches evaluation methods and supports the precise quantification of space quality. Additionally, the study reveals the indirect impact mechanism of space quality on tourist experience through the dimensions of cognition, affect, and behavior, deepening theoretical insights into their relationship. These findings hold significant implications for improving space quality and optimizing tourist experiences in the preservation and renewal of historic districts.

This study, centered on the Nanchang Wanshou Palace, presents findings and methodologies with broader applicability, offering valuable insights for the preservation and development of other historic districts. The balance between cultural heritage conservation and commercial development, a common challenge for many historic areas, can be addressed through the adaptation of these strategies to local contexts. Furthermore, the spatial quality indicators and visitor experience dimensions proposed in this research exhibit wide relevance, contributing to enhanced public engagement and recognition of heritage preservation efforts. Additionally, tools and methodologies such as spatial analysis and eye-tracking, once tailored to specific conditions, can effectively support localized conservation and development initiatives. However, given the diverse cultural, socioeconomic, and policy contexts of various historic districts, the practical implementation of these approaches requires careful contextualization and refinement.

However, certain limitations remain. The timeliness of the data needs improvement, the accuracy of semantic segmentation requires enhancement, the VR eye-tracking experiment requires the participation of volunteers from diverse backgrounds to ensure comprehensive and unbiased results, and the coverage of research indicators should be expanded. Future research could delve deeper into the direct influence mechanisms of historic districts on tourist experience, with a particular focus on cultural symbols and historical ambiance. Developing refined semantic segmentation models tailored for historic districts could improve the accuracy of micro-level element recognition. Additionally, incorporating participants from diverse backgrounds, such as tourists and local residents, into the experiment is essential. These research directions will contribute to the construction of a more comprehensive evaluation system for space quality in historic districts, providing more precise decision-making support for their preservation and renewal, and fostering theoretical and practical innovations in the sustainable development of historic districts.