Next Article in Journal
Development of Safety Domain Ontology Knowledge Base for Fall Accidents
Previous Article in Journal
Research on BIM Technology of Green Building Based on GBSWARE Software
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Multi-Source Data and Semantic Segmentation: Spatial Quality Assessment and Enhancement Strategies for Jinan Mingfu City from a Tourist Perception Perspective

School of Architecture and Urban Planning, Shandong Jianzhu University, Jinan 250101, China
*
Author to whom correspondence should be addressed.
Buildings 2025, 15(13), 2298; https://doi.org/10.3390/buildings15132298
Submission received: 30 May 2025 / Revised: 27 June 2025 / Accepted: 28 June 2025 / Published: 30 June 2025
(This article belongs to the Section Architectural Design, Urban Science, and Real Estate)

Abstract

In the context of cultural tourism integration, tourists’ spatial perception intention is an important carrier of spatial evaluation. In historic cultural districts represented by Jinan Mingfu City, tourists’ perceptual depth remains underexplored, leading to a misalignment between cultural tourism development and spatial quality needs. Taking Jinan Mingfu City as a representative case of a historic cultural district, while the living heritage model has revitalized local economies, the absence of a tourist perspective has resulted in misalignment between cultural tourism development and spatial quality requirements. This study establishes a technical framework encompassing “data crawling-factor aggregation-human-machine collaborative optimization”. It integrates Python web crawlers, SnowNLP sentiment analysis, and TF-IDF text mining technologies to extract physical elements; constructs a three-dimensional evaluation framework of “visual perception-spatial comfort-cultural experience” through SPSS principal component analysis; and quantifies physical element indicators such as green vision rate and signboard clutter index through street view semantic segmentation (OneFormer framework). A synergistic mechanism of machine scoring and manual double-blind scoring is adopted for correlation analysis to determine the impact degree of indicators and optimization strategies. This study identified that indicators such as green vision rate, shading facility coverage, and street enclosure ratio significantly influence tourist evaluations, with a severe deficiency in cultural spaces. Accordingly, it proposes targeted strategies, including visual landscape optimization, facility layout adjustment, and cultural scenario implementation. By breaking away from traditional qualitative evaluation paradigms, this study provides data-based support for the spatial quality enhancement of historic districts, thereby enabling the transformation of these areas from experience-oriented protection to data-driven intelligent renewal and promoting the sustainable development of cultural tourism.

1. Introduction

Under the context of deep cultural tourism integration, the spatial experience quality of cultural destinations has emerged as a critical determinant of tourist satisfaction [1]. Tourist perceptual imagery at destinations profoundly influences travel decision-making, consumption motivation, cultural identity formation, and recommendation intention [2]. This dynamic process, driven by perceived images and ultimately condensed into experience quality, is of vital importance to the development strategy of urban tourism [3]. Driven by the policy of “Cultivating Tourism through Culture and Showcasing Culture through Tourism,” Jinan Mingfu City—as a National 4A-grade Scenic Area and pivotal cultural landmark of the “Spring City” [4]—has seen a remarkable enhancement in its cultural and tourism appeal. The city’s annual tourist volume surged from 64 million to 116 million visits within a three-year period [5]. Notably, the Mingfu City–Daming Lake precinct, covering only 0.42 km2, accommodates 17.2% of the city’s total tourist flow [6], with its single-day reception during the 2025 May Day holiday reaching 3.93 million tourists [7]. This exponential growth in visitation has precipitated massive volumes of real-time perception data, establishing an unprecedented empirical foundation for systematically analyzing tourists’ perceptual imagery towards spatial environments, cultural atmospheres, and service facilities [8]. Current conservation policies for historical blocks predominantly emphasize physical fabric and scholarly perspectives (as seen in research by Lv [9] and Zhao [10]), yet lack a quantitative evaluation and response mechanism for spatial quality based on tourist perception data, this study focuses on the tourist perspective to explore the optimization path of spatial quality in historical blocks.
Tourists, as the core experiencers and evaluation subjects of cultural tourism value in historical blocks [11], have spatial perception serving as a key indicator for measuring the effectiveness of block vitalization [12]. In the field of theoretical research on tourists’ spatial perception, Urry’s [13] ‘The Tourist Gaze’ theory emphasizes the initiative of tourist interpretation; Ai et al. [14] established a tourist perception evaluation system; Ma [15] explored the relationship between tourism experience and tourist satisfaction; Stankov, U. et al. [16] conducted empirical research further proving that social media comments can effectively capture tourists’ embodied perceptions. In terms of technological application, text mining and sentiment analysis have emerged as core methodologies for analyzing tourist-generated content (TGC), demonstrating empirical efficacy in representing collective tourist perceptions. The evolution of text mining technology has traversed critical developmental phases: Manning, C.D. et al. [17] established the theoretical foundations for structured data extraction from textual sources and analyzed interactive models between users and web interfaces; Hearst, M. A. [18] promoted the evolution from “script-based crawlers” to “systematic crawlers”; Rilof, E. et al. [19] pioneered foundational research; Hatzivassiloglou, V. et al. [20] initiated the construction of core sentiment lexicons and affective classification frameworks. The field concurrently witnessed expanding applications, exemplified by Alonso, M. A. et al. [21] analyzing the polarity–intensity dimensions of disinformation effects. Domestically, Li et al. [22] employed SnowNLP for Weibo comment sentiment analysis, while Peng et al. [23] integrated sentiment analytics with LDA modeling to construct a Sentiment-Aware Collaborative Filtering algorithm. These established technical frameworks have been extensively deployed across domains, including destination image perception [12], service quality bench-marking, and experience value identification [14], thereby furnishing robust analytical instruments for capturing tourist perceptions.
In quantitative spatial quality assessments of historical blocks, multi-source data fusion methodologies—spanning Space Syntax and social media sentiment data—have catalyzed a paradigm shift toward quantified research frameworks. For example, Karimi, K. [24] coupled spatial syntax (integration and choice) with social media sentiment data, revealing the impact of street topological structures on tourist pleasure. Maniei, H. et al. [25] demonstrated enhanced pedestrian space safety and sociability potential in traditional alleyway circuits through integrated Space Syntax and behavioral mapping. Concurrently, the integration of streetscape imagery and machine learning technologies has pioneered novel pathways for quantifying spatial quality. For example, Ye et al. [26] used machine learning to calculate the greenery visibility from street view images and conducted an integrated analysis with street accessibility. Wang et al. [27] combined image data with deep learning techniques to establish the relationship between street view elements and perceptual evaluation values. Wang [28] constructed a PSPNet-based semantic segmentation model, combining POI data to quantify street cultural perception and achieve an objective evaluation of spatial quality. Huang et al. [29] measured the correlation between green vision rate and public satisfaction through street view semantic segmentation, within quantitative metrics of spatial quality. Dai et al. [30] proposed a composite index of “green vision rate–street openness–interface enclosure” (weighted 4:3:3), which comprehensively evaluates comfort and optimizes facility layout by integrating tourist review data. Zheng et al. [31] analyzed streetscape data using AI and proposed a green visual index, sky openness index, and color atmosphere index, which are used to quantify the visual coordination between the natural and artificial environment of historic districts. Gu [32] focused on the cultural dimension, introducing the Heritage Conservation Impact Degree and Cultural Element Diversity index. Fang et al. [33] analyzed the impact of historic district texture on street aesthetics through indicators such as Street Height–Width Ratio and Store Signage Count. Han et al. [34] adopted indicators, including Visual Entropy and the Color Richness Index, to assess street comfort in historic cultural districts.
Collectively, these studies converge on an integrated paradigm, synthesizing tourist perception analytics, computational data processing, and multi-dimensional spatial performance indicators. This corpus provides critical theoretical grounding, methodological frameworks, and technical infrastructures for the evidence-based optimization of spatial quality in historical blocks through the tourist perspective.
Two persistent limitations remain: First, standardized analytical frameworks for multi-source data synergy are underdeveloped. Prevailing studies predominantly employ singular data sources, failing to establish correlative analytics between affective semantics and physical spatial elements. Affective semantic data often remain confined to qualitative descriptions, exhibiting insufficient integration with quantified metrics such as the Green View Index and Interface Enclosure Ratio. Second, the collaborative optimization mechanism for multi-source data is not yet mature, where technical tools such as Space Syntax and street view semantic segmentation remain weakly coupled with tourist evaluation frameworks, impeding precise identification of spatial quality thresholds. Addressing these gaps, this research adopts Jinan Mingfu City as an empirical case study. Targeting its high-traffic yet suboptimal experiential outcomes, it integrates Python web crawlers, SnowNLP sentiment analysis, and OneFormer semantic segmentation into a “sentiment element extraction-physical space quantification-indicators impact analysis” workflow, establishing pixel-level associations between tourists’ affective orientations towards material elements and 150 physical elements, thereby resolving the subject–object data dichotomy prevalent in conventional research. A double-blind scoring mechanism involving urban planners and tourists, validated via Spearman’s rank correlation analysis, to establish statistically validated key performance indicators across three dimensions: visual perception–spatial comfort–cultural experience, providing operational quantitative standards for spatial quality evaluation. By synthesizing multi-source data and quantitative analysis, this study constructs a living heritage-aligned spatial quality evaluation system, offering a scientific basis for the sustainable development of historic districts under cultural tourism integration and fostering dynamic equilibrium among protection, development, and tourist needs.

2. Materials and Methods

2.1. Spatial Characteristics and Cultural Heritage Distribution of Jinan Mingfu City

Jinan Mingfu City, situated in Lixia District, Jinan City, Shandong Province, constitutes a vital component of Jinan’s historic urban core and serves as a key area of Jinan’s historical and cultural city. Enclosed by the Moat and Daming Lake, it covers a total area of 3.2 km2 [35]. The city features two quintessential historical and cultural blocks: Jiang Junmiao historical blocks and Fu Rong Street–Bai Huazhou historical blocks. These blocks preserve the majority of the district’s heritage buildings, spring channels, and cultural landscapes, retaining the architectural features and spatial configurations of the Ming and Qing dynasties (Figure 1).
  • Fu Rong Street–Bai Huazhou historical blocks (bounded by Quancheng Road to the south, Daminghu Road to the north, Zhenchi Street to the east, and Gong Yuan qianggen Street to the west) exemplify Jinan’s traditional commercial culture and folk customs, hosting clusters of time-honored eateries, handicraft stores, and cultural and creative boutiques. Baihuazhou, with its natural landscapes and historical and cultural heritage, has become a premier destination for tourists to experience Jinan’s spring water culture and historic urban ambiance.
  • Jiang Junmiao historical blocks (delimited by Quancheng Road to the south, Daminghu Road to the north, Bianzhi Lane to the east, and Taiping Temple Street and Xichenggen Street to the west) are distinguished by their religious heritage and historical architecture, including landmarks such as the Guandi Temple and Fu Xue Confucian Temple. These ancient buildings not only showcase the artistic style of Jinan’s ancient architecture but also reflect the city’s profound historical and cultural heritage.
To comprehensively document the historical and cultural resources of Mingfu City, a historical resource database was constructed. The database encompasses distribution information of heritage conservation units and intangible cultural heritage. Through collecting and organizing relevant materials, detailed records of cultural relics protection units within Mingfu City were meticulously documented. These records include attributes such as name, location, historical period, architectural style, and other pertinent information. Meanwhile, intangible cultural heritage sites were cataloged, with metadata recorded on aspects such as name and type. Spatial analysis of this dataset enables a clear understanding of the distribution patterns and conservation status of heritage resources, providing a scientific basis for subsequent research and conservation work (Figure 2).
  • Heritage architecture: A total of 276 traditional buildings, including 11 provincial-level conservation units such as Fu Xue Confucian Temple, Chen Mian’s Zhuangyuan Mansion, and 30 municipal-level conservation units such as Lao She’s Former Residence, Hou Zaimen Christian Church;
  • Intangible cultural heritage sites: National-level intangible cultural heritage sites such as Jinan Shadow Puppetry Theater, Lu Embroidery, Cloisonné Workshops, and the birthplace of Qushui Liushang culture;
  • Characteristic spaces: The spring–alley–courtyard spatial fabric, such as Furong Spring, Wangfu Pool, Ming-Qing dynasty streets and alleys, such as Qushuiting Street, Jiang Junmiao Street, and spring water folk activities, such as the Lotus Festival and Dragon Lantern Fair.

2.2. Research Framework

This study employs Python 3.8.7 web crawlers to collect 9000 tourist reviews (2019–2023) from platforms including Mafengwo, Xiaohongshu, and Ctrip. Emotional word segments were extracted via the SnowNLP sentiment analysis module. A sentiment intensity matrix was generated through −3 to 3 standardization processing. This quantifies the public’s emotional tendencies (positive/negative/neutral) toward high-frequency physical elements in the block, such as spring water, historic streets, and architecture. Concurrently, TF-IDF text mining identifies core semantic features. Through SPSS principal component analysis (KMO = 0.85, Bartlett’s sphericity test p < 0.001) with maximum variance rotation, the vocabulary was merged into three dimensions: visual perception, spatial comfort, and cultural experience. Building on literature reviews and expert consultations, an evaluation framework was formulated:
  • Visual dimension: Prioritizes aesthetic coherence;
  • Comfort dimension: Focuses on spatial accessibility and facility convenience;
  • Cultural dimension: Emphasizes historical symbol preservation and narrative immersion.
An initial index pool containing 10 observed variables was constructed, with calculation algorithms further derived for each indicator. Based on Space Syntax theory, DepthmapX 10-0.6.0 software was used to calculate the integration value and connectivity value of 36 main streets in Mingfu City. Combined with Heritage Visibility Density (HVD), five typical streets were selected as study samples. Using ArcMap, 71 sampling points were set at 30 m intervals to systematically capture Baidu Street View panoramic images. The OneFormer semantic segmentation framework (pre-trained on the ADE20K dataset) was employed for pixel-level analysis, quantifying proportional distributions of 150 physical elements, such as buildings, vegetation, and signage. This approach transforms abstract concepts, such as the signboard clutter index and street enclosure ratio, into computable metrics, overcoming the subjectivity inherent in manual surveys.
To ensure the objectivity and reliability of evaluation results, a human–machine collaborative evaluation mechanism was implemented: Automated scoring via semantic analysis and element proportion algorithms, and double-blind human evaluations by experts and tourists. Through Spearman rank correlation analysis, the quantitative correlations between machine indicators and manual scores were revealed (Figure 3).

3. Data and Methodology

3.1. Street Potential Assessment and Screening Based on Space Syntax

To obtain a detailed block network map of the study area, the research first utilized OSM (OpenStreetMap, https://www.openstreetmap.org, accessed on 17 December 2024) road network data based on 2024 China’s administrative divisions to acquire the initial road network of Jinan Mingfu City. Subsequently, the initial road network was refined using historical district conservation blueprints released by the Jinan Municipal Bureau of Resources, and a CAD road network map was drafted using AutoCAD 2014 software. Considering the small-scale characteristics of streets in historical blocks, this study defined road intersections as the basis for street segmentation. After completing the CAD road network map, it was imported into DepthmapX software for further analysis (Figure 4 and Figure 5).
In the geographic research progress of spatial syntax, Luo [36] and Tao Wei explored the relationship between the built environment and the pedestrian path selection behavior, which indicates that the shortest path and the topological path are the two most discussed indicators. Consequently, the integration value and connectivity value were selected as analytical indicators within the Space Syntax framework. The integration value reflects the accessibility of a space within a region by analyzing the topological relationships and connectivity patterns among spatial elements [37]. It serves as an indicator of spatial centrality and aggregation [38]. The connectivity value refers to the frequency of the shortest topological distance connecting two independent nodes in the spatial system [39]. It represents the collection of the shortest topological distances from any element to others in the system, indicating the likelihood of a spatial unit being traversed by other shortest paths [40].
Regarding the determination of high and low value ranges for indicators, scholars’ criteria vary. Typically, based on the color assignment in DepthmapX visualization results, warm colors (red/yellow) represent high values, while cool colors (blue) represent low values [39,41,42,43]. Some studies set percentage thresholds (e.g., top 25% [44], 20% [45], or 10% [46]), but there is no unified standard. More research papers adopt the mean value as the criterion for distinguishing high and low levels [47,48]. In the numerical calculations of this study area, the values are relatively small, making the mean method suitable as the basis for judgment. Furthermore, due to the differing numerical scales of integration value and connectivity value, normalization processing (to the [0–1] interval) was performed, resulting in a mean integration value of 0.5631 and a mean connectivity value of 0.4482.
By evaluating two key spatial syntax metrics—the integration value and connectivity value—the streets of Mingfu City were classified into four categories:
  • High-efficiency streets (In > 0.5631, Cn > 0.4482): High integration value and connectivity value, serving as the core streets with significant cultural tourism potential.
  • Potential streets (In > 0.5631, Cn ≤ 0.4482): High connectivity value but low integration value. These streets could be transformed into culturally valuable corridors through targeted improvements.
  • Isolated streets (In ≤ 0.5631, Cn ≤ 0.4482): Low integration value and connectivity value, largely disconnected from the main street network.
  • Transitional streets (In ≤ 0.5631, Cn > 0.4482): Intermediate integration value and connectivity value, exhibiting moderate cultural tourism potential.
In historic districts, the planning of internal road networks requires a balanced approach that harmonizes heritage conservation with spatial–functional coordination. Simply pursuing topological indicators such as integration value and connectivity value from Space Syntax is not applicable to such special regions. Due to the presence of numerous protected heritage assets within the block, solely pursuing high integration values and connectivity values may lead to heritage damage. Therefore, by overlaying the Heritage Visibility Density (HVD), the streets were further classified for conservation, enabling the proposal of targeted protection measures. The street potential assessment methodology based on Space Syntax integrates the integration value and connectivity value through spatial overlay analysis (Figure 6). Combined with Heritage Visibility Density (HVD), to construct a Composite Potential Index (CPI). Five typical streets were subsequently identified for in-depth analysis.
The HVD is defined as the number of heritage assets (historic buildings, inscriptions, etc.) within every 100 m of the visual field, which is identified and counted via semantic segmentation of heritage-related labels (Table 1). (See details in the Published List of Provincial-Level Cultural Relics Protection Units in Shandong Province http://whhly.shandong.gov.cn/art/2023/6/8/art_315161_10323830.html, accessed on 23 December 2024 and the Directory of Municipal-Level Cultural Relics Protection Units in Jinan City https://data.sd.gov.cn/portal/catalog/d132a73610c74f2495146964ee0e721f, accessed on 25 December 2024).
Heritage   Visibility   Density = Number   of   Cultural   Heritage   Sites View   Length   ( in   hundreds   of   meters )
The CPI is calculated through weighted integration value (In), connectivity value (Cn), and the HVD using the formula:
CPI = 0.4 × In + 0.3 × Cn + 0.3 × HVD
Based on the overlay analysis of the CPI, Space Syntax metrics (integration value and connectivity value), and the HVD, five streets (Qu Shuiting Street, Shuang Zhongci Street, Jiang Junmiao Street, Hou Zaimen Street, and Shengfu Front Street) were selected as typical samples for further analysis. These streets encompass the primary street typologies of high-efficiency, isolated, and potential types (with no transitional streets included in these historic districts). They exhibit significant variations across CPI values (ranging from 0.327 to 2.067), heritage density (0 to 5.2 sites/100 m), and spatial topological characteristics. This diversity ensures a comprehensive reflection of the distinct spatial features and problem types present in the street network of the study area, thereby guaranteeing the typicality and persuasiveness of the case selection. Furthermore, considering the workload associated with field surveys and multi-source data collection, selecting five streets enhances the operational feasibility of this study while maintaining representativeness. The specific characteristics of these streets are detailed in Table 2.

3.2. Street View Image Data Acquisition

In conducting street view data collection for the five selected streets, the sampling point configuration critically determines data representativeness and analytical accuracy. For high-efficiency streets, a 30 m equidistant sampling interval was implemented. For isolated streets with HVD, a dense sampling strategy was adopted to holistically capture spatial characteristics and support data-driven interventions for enhancing tourism attractiveness. This dual-strategy framework yielded 71 sampling points across the study area.
Leveraging the Baidu Map API endpoint, Python crawlers systematically retrieved Baidu Street View images, each embedded with metadata including size, location, heading, and pitch. At each sampling point, street view images were captured in four directions (0°, 90°, 180°, and 270°) (Figure 7 and Figure 8), resulting in 284 validated street view images for subsequent analysis.

3.3. Public Perception Element Extraction

With the rapid development of Web 3.0 [49], social media platforms have become a crucial channel for tourists to express experiential perceptions [50]. Research has proven that TGC from travel review websites holds significant value for the exploration and analysis of destination image perception [51]. Most scholars prioritize platforms such as Ctrip, Qunar, and Mafengwo as key sources for user reviews [3,52,53,54,55,56]. To investigate tourists’ perceptions and experiences in Jinan Mingfu City, a multi-source data acquisition system was set up, encompassing major cultural tourism social platforms such as Xiaohongshu, Qunar, Mafengwo, and Weibo. Utilizing Python web crawlers and the Octopus 8 software for targeted and collected review data, 8985 raw data entries were collected, including user profiles, textual descriptions, image links, and rating details. During the data preprocessing phase, duplicate reviews (127 entries), invalid data (215 entries unrelated to Mingfu City), and low-quality texts (389 entries with fewer than 10 characters) were sequentially removed to form the foundational corpus.
The research referred to the spatial element classification system of Jinan Mingfu City proposed by Li [57] and combined the collation of historic district material environmental components by Du [58], Zhao [59], etc., to summarize a material element keyword repository specific to Mingfu City, covering 136 core terms such as “spring water”, “park”, “lakeside”, “lotus”, and “architecture”. Early applications of SnowNLP sentiment analysis used a scoring range of 0–1 [60,61]. As technology advanced, an increasing number of studies began to further refine the granularity of sentiment scores. Scholars such as He et al. [53] and Luo et al. [56] utilized the sentiment analysis functionality of RostCM6, introducing three intensity levels—neutral, moderate, and high—within the categories of positive and negative sentiment. Zheng [62] categorized sentiment into evaluative positivity, evaluative attitude, evaluative intensity, and evaluative focus in her thesis. To achieve more specific emotional tendencies for physical elements, the Python code was enhanced to incorporate extremely strong emotional biases, implemented through adverbial–noun combinations, assigning sentiment scores of +3 and −3.
In the sentiment semantic analysis phase, a Python sentiment sentence processing model (Material_Environment_Classifier) was first developed to filter 15,648 valid sentiment text segments. Subsequently, a dual-dimensional sentiment polarity–intensity model was constructed using natural language processing (NLP) toolkits (NLTK [63] + SnowNLP). This involved the following steps:
  • Sentence segmentation and tokenization of preprocessed texts.
Sentiment lexicon tagging based on the HowNet Sentiment Lexicon to identify affective lexemes.
  • Weighted scoring algorithm assigning sentiment scores for each review:
    Positive: +3 to +1
    Negative: −1 to −3
    Neutral: 0
A theme-element coupling extraction method was then employed, utilizing the TF-IDF algorithm to perform term frequency statistics on material element keywords. Low-frequency irrelevant terms were filtered using mutual information (MI) and the chi-square test (χ2) [64]. Through a cross-analysis of sentiment scores and element term frequencies, the correlation strength between material elements and emotional tendencies in tourist experiences was revealed (Table 3 and Table 4).

4. Results

4.1. Development of a Three-Dimensional Indicator Framework

Drawing upon the research dimensions established by scholars in the spatial quality conservation of historic districts, the 136 physical elements were categorized and processed. Principal component analysis (PCA) in SPSS was employed to extract key features. For instance, Zhang [65] emphasized the need for a synergistic relationship among physical spatial morphology, functional adaptability, and cultural value continuity in historic urban conservation. Li et al. [66] highlighted the significance of the “natural landscapes–historic streets and alleys–cultural activities” triad in their study of the Ciqikou Historic District. Wang, Y. et al. [67] based on cultural landscape gene theory, analyzed the perceptual differences between tourists and residents across the dimensions of “material landscapes–human actions–place meanings.” Jiang, S. Y. et al. [68] utilized street view image data to evaluate historic district spatial quality from the three dimensions of “vitality, safety, and landscape.” Building on the research of these scholars, core variables—including total word frequency, mean sentiment score, positive frequency, negative frequency, and sentiment variance—were extracted from the 136 physical element terms derived from word frequency statistics and sentiment scoring, thereby constructing a multi-dimensional dataset.
A multi-dimensional dataset was constructed by extracting core variables—including total word frequency, mean sentiment score, positive/negative frequency, and sentiment variance—from the 136 material element keywords derived through word frequency statistics and sentiment scoring (Table 5). Principal component analysis (PCA) was subsequently conducted using SPSS 25 software [69], following these standardized procedures: First, the original variables were standardized to eliminate dimensional differences; the suitability of the data for factor analysis was confirmed through the KMO test (KMO = 0.82) and Bartlett’s sphericity test [70] (p < 0.001); then, the principal component method was used to extract common factors with eigenvalues greater than 1, resulting in three principal components with a cumulative variance explanation rate of 68.72%; finally, varimax rotation was applied to obtain a clear factor loading matrix.
Based on the rotated factor loading matrix, variables were categorized according to their loading coefficients (absolute value > 0.4) across principal components (Table 6).
The first principal component (Factor 1) was characterized by significantly higher loading values from visual-related elements such as “spring water” (5.038), “park” (3.997), “lotus” (2.526), and “architecture” (0.787). These variables collectively reflect the visual attributes of the physical environment, including color, form, vegetation, and water bodies. This dimension is classified as the “visual perception” dimension.
The second principal component (Factor 2) emphasized spatial amenity indicators such as “lively atmosphere” (2.272) and “leisure” (1.026), as well as facility-related terms such as “hutong” (2.000), “stalls” (1.850), and “stone bench” (−1.281) highlighting functional convenience, physiological comfort, and pedestrian activity patterns, and these are categorized as the “spatial comfort” dimension.
The third principal component (Factor 3), historical architecture-related terms, such as “Huixianlou Restaurant” (2.052) and “Lao She’s Former Residence” (1.039), exhibit strong correlations with intangible cultural practices such as “listen to crosstalk” (1.669) and “folk performance” (1.656). This cluster focuses on the cultural symbol-bearing and living experience of historical sites, thus being classified as the “cultural experience” dimension.
Based on Li’s “Spring-Alley-Courtyard” contextual framework for Mingfu City [57] and integrating Du “Resident-Tourist Symbiosis” theory [58] with domestic and international research on street space quality, ten core indicators were systematically selected across three dimensions: visual perception, spatial comfort, and cultural experience, with the indicators closely linked to the factors derived from the three principal components (Table 7 and Table 8).
  • The visual dimension builds upon Fang, Z.G.’s quantitative research on street aesthetic indicators [33], focusing on the visual attributes of the physical environment impacting tourist experience: green vision rate, sky visibility rate, and water visibility ratio. Additionally, the signboard clutter index was introduced to reflect conflicts between commercial interfaces and historic landscapes. Special terms such as “Spring Water,” “Lakeside,” “Pearl Spring,” and “Stream” in Factor 1 are rarely found in other scholars’ indicators because of the spring city characteristics of Jinan Mingfu City. Therefore, the water visibility ratio was added.
  • The spatial comfort dimension is bench-marked against Han, J.W.’s “Street Comfort” model [34], quantifying the effective pedestrian clear width, shading facility coverage, and street enclosure ratio. Terms such as “stone bench” and “bench” in Factor 2 indicate tourists’ emphasis on resting spaces. Thus, the public seating density indicator was added to quantify resting convenience.
  • The cultural dimension integrates Gu, W.Q.’s “Cultural Heritage Transmission Pathway” theory [32]. Combining cultural activity spaces such as “listen to crosstalk,” “folk performance,” “Lotus Festival,” and “market” identified in Factor 3, indicators for the cultural activity space ratio and informal vending space occupancy were established, translating “Intangible Cultural Heritage (ICH) Activation” into measurable street vitality indicators.
Specific indicators and computational methods are as follows:
Table 7. Indicator definitions and computational methods.
Table 7. Indicator definitions and computational methods.
Dimension IndicatorScholarConceptual DefinitionAlgorithmic Formulation
Visual DimensionGreen Vision RateFang, Z.G. [33]
Huang, J.X. [34]
Zheng, Y. [31]
Pei, Y. [71]
Pixel share of vegetation (trees, grass) in the field of view and identification of vegetation areas by semantic segmentation. G r e e n   V i s i o n   R a t e = V e g e t a t i o n   P i x e l   C o u n t T o t a l   P i x e l   C o u n t × 100 %
Sky Visibility RateFang, Z.G. [33]
Yang, J.Y. [72]
Han, J.W. [34]
Sky visible area share, reflecting the sense of spatial oppression, and semantic segmentation to extract sky pixels. Sky   Visibility   Rate = S k y   P i x e l   C o u n t T o t a l   P i x e l   C o u n t × 100 %
Water Visibility Ratio Fang, Z.G. [33]
Xiong, X [73]
The proportion of water bodies, such as springs and rivers, that are visible in the streetscape, and semantic segmentation to recognize water body pixels. W a t e r   V i s i b i l i t y   R a t i o = Water   Body   Pixel   Count Total   Pixel   Count × 100 %
Signboard Clutter IndexFang, Z.G. [33]
He, Z.Y. [74]
Percentage of commercial advertisements with non-historical and cultural elements, semantic segmentation identifies billboard pixels, and calculates density. S i g n b o a r d   C l u t t e r   I n d e x = A d v e r t i s i n g   S i g n   P i x e l   C o u n t T o t a l   P i x e l   C o u n t × 100 %
Comfort DimensionEffective Pedestrian Clear WidthHan, J.W. [34]
Zhang, Y.C. [75]
Percentage of passable width after subtracting obstacles, semantic segmentation identifies occupying facilities (electrical boxes, stalls), and calculates the available space. E f f e c t i v e   P e d e s t r i a n   C l e a r   W i d t h = Usable   Width   Pixel   Count Total   Width   Pixel   Count × 100 %
Shading Facility CoverageHan, J.W. [74]
Feng, X.X. [76]
Percentage of shaded areas, such as pergola, shade, etc., and semantic segmentation to identify shaded areas (shadows + amenities). S h a d i n g   F a c i l i t y   C o v e r a g e = Shade   Area   Pixel   Count Total   Pixel   Count × 100 %
Public Seating DensityHan, J.W. [34]
Fang, H.Y. [77]
Percentage of street seats, etc., semantic segmentation identifies seats and counts them.Public Seating Density =
Number   of   Seats Street   Length   ( in   hundreds   of   meters ) × 100 %
Street Enclosure RatioHan, J.W. [34]
Fang, Z.G. [33]
Evaluate the degree of interface enclosure and street openness, composed of elements such as walls, buildings, and pavements, affecting the privacy and openness of the space. Street   Enclosure   Ratio = P w a l l + P b u i l d i n g P s i d e w a l k × 100 %
Cultural DimensionCultural Activity Space RatioGu, W.Q. [32]
Jiang, L. [78]
Proportion of public space available for cultural activities (plazas, stages, bazaars) to the total area of the neighborhood, and semantic segmentation to identify activity spaces.Cultural Activity Space Ratio =
Activity   Space   Pixel   Count Total   Pixel   Count × 100 %
Informal Vending Space OccupancyGu, W.Q. [32]
Fang, Z.G. [33]
Proportion of pavement area occupied by temporary stalls (snacks, handicrafts) in the street, semantic segmentation to identify stall areas.Informal Vending Space Occupancy =
Stall   Area   Pixel   Count Sidewalk   Area   Pixel   Count × 100 %

4.2. Machine Scoring and Human Scoring

Semantic segmentation, a deep learning technique, enables pixel-level recognition and delineation of distinct objects and scenes within images. Employing a fully convolutional network (FCN-8s) trained on the ADE20K dataset, street view images undergo semantic segmentation to classify pixels into 150 urban elements, including road, sky, and car (Figure 9). For each sampling point, the pixel ratio of a specific element is calculated by averaging its proportion across four-direction street view images (0°, 90°, 180°, and 270°), yielding the final pixel ratio of that element at the sampling point.
Through deep learning semantic segmentation technology, pixel-level analysis of street view images is conducted. Based on indicator calculation formulas, indicators such as the green vision rate and signboard clutter index are quantified (Figure 10). The scatter plots of the proportion of each indicator are shown as follows (Figure 11):
Machine scoring was derived by normalizing semantic segmentation-derived metric percentages through min–max linear scaling, where the observed range of each metric (minimum to maximum percentage) was proportionally mapped to a standardized interval of [−3, +3], yielding the machine-calculated scores for individual indicators (Figure 12).
To enhance the objectivity and scientific rigor of spatial quality assessment in Jinan Mingfu City historic district, this study introduces a “double-blind manual scoring” mechanism, integrating the perspectives of planning experts and tourists to conduct a refined assessment of the three dimensions of “Visual Perception-Spatial Comfort-Cultural Experience” through street view images as the medium.
  • Methodological Protocol:
Data Acquisition: High-definition streets view images of five representative streets were captured via big data crawling, constructing a standardized image repository comprising a total of over 71 images. To facilitate comparison of scoring data, the images were numbered from 0 to 70, based on sampling points.
  • Scoring Process:
Expert Group: Including 10 certified urban planners.
Visitor Group: Including 50 participants (local residents and non-local tourists).
Double-Blind Design: Participants independently scored randomly assigned images through a uniform scoring interface without knowledge of the evaluation targets or each other’s identities, effectively mitigating subjective biases arising from regional preferences or professional standpoints.
The manual scoring results are shown in Figure 13 and Figure 14:
These data indicate that expert scores across all three dimensions of the historic district were consistently higher than those of tourists (Table 8). This highlights that current protection and development policies overly focus on the physical spatial ontology and expert perspective [9,10], lacking authentic tourist experience data. Consequently, it is difficult to enhance the tourist-perceived spatial quality.
Table 8. Mean scores of expert and tourist ratings across three dimensions.
Table 8. Mean scores of expert and tourist ratings across three dimensions.
DimensionExpert GroupVisitor GroupDifferentiation Characteristics
Visual Dimension1.281.18Expert scores are slightly higher (+8.5%)
Comfort Dimension1.120.86Expert scores are significantly higher (+30%)
Cultural Dimension1.050.64Expert scores are substantially higher (+64%)
The visual perception dimension exhibited high consistency (e.g., Qu Shuiting Street experts 2.74 versus tourists 2.67), with negligible ambiguity. However, significant differences emerged in the cultural experience dimension, exemplified by Jiang Junmiao Street (experts 1.35 vs. tourists 0.18, a 650% difference) and Shuang Zhongci Street (experts 0.68 vs. tourists 0.26, a 161% difference). This discrepancy stems from experts prioritizing architectural details (e.g., decorated archways and brick carvings) and traditional activities, whereas tourists emphasize overall cleanliness, bustling atmosphere, and comfort (influenced by pedestrian flow), coupled with insufficient cultural symbol cognition. Therefore, this study will integrate the dual perspectives of experts and tourists for correlation analysis of the machine scoring, ensuring the scoring results reflect both professional conservation requirements and capture the tourists’ spatial perception experience.

4.3. Analysis of Indicator Weighting Adjustment and Data Calibration Based on Human–Machine Scoring Correlations

Data format consistency between machine-generated scores and double-blind human evaluations was verified [80]. Rigorous validation confirmed alignment in sample IDs (ranging from 0 to 70) and street groupings, including Shengfu Front Street, Hou Zaimen Street, etc. Regarding data normalization, both human ratings (visual/comfort/cultural dimensions) and machine scores (10 indicators) employed a unified [−3, +3] scale, eliminating the necessity for additional normalization. Employing Spearman’s rank correlation analysis [81], combined with effect size (Cohen’s d) [82] to determine the magnitude of impact, this study analyzed the associations between the ten machine-evaluated indicators and the three human-rated dimensions (Figure 15 and Table 9):
  • Visual perception dimension: The green vision rate and shading facility coverage demonstrated strong positive correlations with visual perception scores, with correlation coefficients of 0.656 and 0.671. This indicates that vegetation coverage and shading infrastructure critically enhance visual experiences in urban streets and parks. Conversely, the street enclosure ratio exhibited a strong negative correlation (−0.604), suggesting excessive enclosure induces visual oppressiveness and necessitates calibrated enclosure control in urban design. Indicators such as the signboard clutter index, cultural activity space ratio, and informal vending space occupancy showed negligible correlations (|r| < 0.1), indicating minimal impact on visual evaluation.
  • Spatial comfort dimension: The green vision rate and shading facility coverage maintained moderate positive correlations with comfort scores, with correlation coefficients of 0.442 and 0.485. The street enclosure ratio has a negative correlation (−0.469) with comfort evaluation scores, meaning that high enclosure is unfavorable for people to obtain a comfortable experience. Indicators such as the water visibility ratio and signboard clutter index exhibited weak associations (|r| < 0.2), implying limited influence on comfort.
  • Cultural experience dimension: Most machine-evaluated indicators displayed statistically insignificant correlations with cultural scores.
The green vision rate, shading facility coverage, and effective pedestrian clear width exhibit large positive effects on visual perception scores (d = 1.28, 1.28, 0.99) and moderate positive effects on spatial comfort scores (d = 0.69, 0.77, 0.70). This indicates that indicators within the visual perception and spatial comfort dimensions demonstrate high consistency between machine-derived and manual evaluations. Such alignment stems from significant physiological and psychological commonalities in human responses to visual and comfort-related stimuli, coupled with the intuitive nature of these indicators, which are readily assessable through direct visual observation and bodily experience. Moreover, they are also amenable to accurate machine quantification via measurable parameters.
In contrast, machine–manual correlations for cultural dimension indicators (e.g., cultural activity space ratio, and informal vending space occupancy) are generally weak. This discrepancy arises primarily from the highly subjective interpretation of cultural connotations, which varies substantially based on individual backgrounds, experiences, and emotions. Furthermore, cultural elements (e.g., historical heritage, folk customs, and artistic ambiance) are inherently complex and often intangible, posing challenges for machine-based capture and quantification (e.g., ordinary corners imbued with historical memory). Additionally, machine scoring typically fails to account for contextuality and narrativity, whereas manual evaluations prioritize emotional resonance embedded in cultural scenarios. Therefore, for the assessment of the cultural experience dimension, it is necessary to integrate manual subjective evaluation (such as questionnaires and expert assessment) based on machine quantification data, and commit to deeper analysis of cultural elements, explore methods for translating intangible culture into quantifiable characteristics, and develop algorithms capable of recognizing contextuality and narrativity to incorporate into the scoring system.

5. Discussion

Based on the three-dimensional evaluation results of “visual perception-spatial comfort-cultural experience” and indicator impact analysis, this study identified that green vision rate, effective pedestrian clear width, shading facility coverage, and street enclosure ratio are the core driving indicators. For these indicators, the mean values across streets were compared (averaging machine-calculated metrics and manual double-blind scores) to obtain indicator proportion and score data for each street (Table 10). This data-driven analysis served as the basis for proposing targeted enhancement strategies.
As shown in Table 11 above, Qu Shuiting Street scored well across all dimensions and exhibited a relatively low street enclosure ratio. In contrast, the other streets demonstrated significant deficiencies in certain aspects. Therefore, to address these issues and consider the substantial disparities observed in the cultural dimension, targeted and highly feasible improvement recommendations are proposed for specific streets across three domains: visual landscape optimization, facility layout adjustment, and cultural scenario implementation. These recommendations aim to enhance the spatial quality enhancement strategies for Jinan Mingfu City’s historic cultural district, thereby achieving a dynamic balance between preservation and revitalization.

5.1. Visual Landscape Optimization Strategies

Shuang Zhongci Street and Jiang Junmiao Street exhibit significant deficiencies in the green vision rate, resulting in monotonous streetscape visuals that lack vitality and aesthetic coherence, thereby failing to deliver a comfortable visual experience. It is recommended to implement strategic greening interventions across roadside areas, building walls, and public spaces along these streets. For example, street trees are planted on both sides of the road, and tree species that are adapted to the local climate and soil conditions, with beautiful shapes and lush branches and leaves, such as Sophora japonica and Ginkgo biloba, are selected; vertical greening is set up on the walls of buildings, and climbing plants such as ivy and creepers are planted to enrich the street facade landscape (Figure 16).

5.2. Facility Layout Adjustment Strategies

Regarding the indicators of shading facility coverage and effective pedestrian clear width, Shengfu Front Street, Hou Zaimen Street, and Shuang Zhongci Street performed poorly. Shuang Zhongci Street exhibited the weakest shading facility coverage (0.047), which makes pedestrians lack shade when walking on the street, especially in hot summer weather, which greatly reduces their comfort. Hou Zaimen Street had the narrowest effective pedestrian clear width (0.047), which is not conducive to the smooth passage of pedestrians and also restricts the development of public activities on the street. Therefore, facility layout adjustment and optimization within the district were undertaken based on practical conditions. For example, widen the sidewalks by appropriately compressing the width of the roadway and adjusting the outdoor space of shops on both sides of the street. In the widening process, pay attention to maintaining the historical style of the street and use paving materials that match the original street materials, such as bluestone slabs and permeable bricks, to improve the comfort and safety of pedestrians. Furthermore, public seating was rationally installed along the sidewalks, and shading facilities were added. Traditional pavilions and pergolas can be used to meet the shade requirements and blend in with the style of the district (Figure 17).

5.3. Cultural Scenario Implementation Strategies

In the cultural experience dimension, although Hou Zaimen Street achieved a relatively high cultural score of 0.97, the scores for Shengfu Front Street (0.534286), Jiang Junmiao Street (0.479444), and Shuang Zhongci Street (0.578667) indicate an overall deficiency in cultural ambiance cultivation. Based on the analysis of cultural indicator correlation differences and tourist experience demands, this study proposes the following cultural scenario implementation strategies to enhance the perception of cultural atmosphere in the district. These strategies primarily aim to deeply strengthen the district’s cultural narrative continuity, construct an innovative ecosystem for the transmission of living heritage, and achieve organic integration of historical-cultural resources with modern cultural tourism, thereby deepening tourists’ cultural perception of the district (Figure 18).
  • Water narrative axis: Taking the Qushui River as the vein, it is proposed to plan a full-process experiential chain of “water drawing-tea tasting-lantern releasing” to recreate the classical literati ambiance of Qushui Liushang. Tourists engage in spring water tea brewing rituals, embodying millennia of spring culture. Then, at nightfall, they can participate in lantern-floating ceremonies, comprehending the poetic charm of Jinan amid sparkling ripples.
  • Alley memory wall: Building upon the historical depth of Jiang Junmiao Street and referencing the implemented example of Fuzhou’s Three Lanes and Seven Alleys (Sanfang Qixiang) Historical and Cultural District AR Metaverse Experience [83,84] (which has achieved digital restoration and display of the Ming-Qing era alley layout), the plan is to use AR (augmented reality) technology to dynamically project scenes of late Qing dynasty market life onto the mottled walls. Tourists can trigger a time-traveling interactive experience by touching the walls, thus making historical scenes “come alive” in the present.
  • Courtyard revitalization: Consider transforming the Jinju Lane into communal hubs for intangible cultural heritage, creating open-access ICH transmission spaces. Regular activities such as cloisonné craftsmanship experiences and Shandong cuisine culinary classes are held, allowing tourists to appreciate the charm of intangible cultural heritage through hands-on practice and promoting the intergenerational inheritance of traditional culture.

6. Conclusions

This study established a technical path of “data crawling, factor aggregation, and human-computer collaborative optimization”. By integrating multi-source data and semantic segmentation technologies, it achieved quantitative assessment and empirical optimization of the spatial quality in Jinan Mingfu City Historic District. The developed three-dimensional indicator system—“visual perception, spatial comfort, and cultural experience”—confirmed that indicators such as the green vision rate, shading facility coverage, and street enclosure ratio significantly influence tourists’ spatial perception, providing operational quantitative standards for spatial quality evaluation in historic districts. This closed-loop model of “data diagnosis–indicator quantification–strategy generation” provides a replicable technical paradigm for transforming historic districts from empirical protection to intelligent renewal.
Although the current research has broken through the limitations of traditional qualitative evaluation, machine learning algorithms still have limitations in capturing tourists’ subjective perception of historical and cultural connotations (e.g., artistic expression of architectural details, emotional memory of the district). Quantifying the cultural experience dimension requires continued integration with the double-blind scoring mechanism. Furthermore, the existing indicator system has limited ability to quantify tourists’ dynamic behaviors (e.g., length of stay and path selection). In the future, GPS trajectory and eye-tracking data can be integrated to improve the evaluation dimensions. Future research should refine the multi-source data collaborative optimization paradigm, such as integrating real-time perception data from eye-tracking and physiological indicators to establish a more comprehensive tourist experience evaluation system. It should also explore the combination of cultural heritage semiotics and place attachment theory with semantic segmentation technology to develop a bidirectional mapping model of “physical space-cultural meaning”, thereby enhancing the quantification capacity for subjective perceptions such as historical ambiance and cultural environment.
This study provides technical support for the transition of historic districts from experience-oriented conservation to data-driven intelligent renewal, helping to realize the dynamic balance between conservation, development, and tourists’ needs in the context of cultural and tourism integration. The methodology can provide quantitative tools for similar districts, such as Nanluogu Lane in Beijing, Pingjiang Road in Suzhou, etc., and help to realize the dynamic balance between cultural heritage preservation and sustainable development of the city.

Author Contributions

Conceptualization, L.C.; methodology, L.C.; software, X.C.; validation, L.C. and X.C.; formal analysis, L.C.; investigation, X.C.; resources, L.C.; data curation, L.C.; writing—original draft preparation, X.C.; writing—review and editing, L.C. and Z.L.; supervision, L.C.; project administration, Z.L.; funding acquisition, Z.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China, grant number 52308025, the Natural Science Foundation of Shandong Province, grant number ZR2022QE293, Shandong Province Key Art Project, grant number ZD202008273, the Youth Innovation Team Program for Higher Educational of Shandong Province, grant number 2023KJ325, and Shandong Provincial Fund of Key Advantageous Discipline (Architecture).

Data Availability Statement

Data presented in this study are available within this article.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
HVDHeritage Visibility Density
CPIComposite Potential Index
NLPNatural Language Processing

References

  1. Guo, W.; Wang, Y.Y.; Hao, R.N. Tourism Experience Quality of Cultural Elements in Revolutionary Heritage Tourism Sites: An Analysis Based on Online Textual Data. Econ. Forum. 2024, 38, 69–84. [Google Scholar]
  2. Afshardoost, M.; Eshaghi, M.S. Destination image and tourist behavioural intentions: A meta-analysis. Tour. Manag. 2020, 81, 104154. [Google Scholar] [CrossRef]
  3. Yan, Y.; Li, B. Identification, Evolution and Asymmetric Influence of Tourists’ Perceived Imagery in Ice and Snow Tourism Cities from the Perspective of Integration of Culture, Sports and Tourism. J. Jilin Sport Univ. 2025, 41, 9–20. [Google Scholar]
  4. The 14th Five-Year Plan for Cultural and Tourism Development. Available online: https://lyj.nanyang.gov.cn/2022/04-01/175324.html (accessed on 16 April 2025).
  5. Shandong Provincial Bureau of Statistics. Available online: http://tjj.shandong.gov.cn/col/col6279/index.html (accessed on 16 April 2025).
  6. Analyzing Tourism Development in Jinan through Ctrip Big Data. Available online: https://baijiahao.baidu.com/s?id=1816565876875339932 (accessed on 20 April 2025).
  7. Jinan Cultural Tourism Review: This May Day Holiday, Jinan Is Turning Heads! Available online: https://mp.weixin.qq.com/s/1r1GJySGn7cv1B2nQ1OIZg (accessed on 20 April 2025).
  8. Huang, W.; Zhou, L.; Zhu, Z.M. Analysis of Audience Satisfaction with On-site Cultural Performances at Tourist Destinations: A Case Study of “Tianmen Fox Fairy” in Zhangjiajie. J. Jishou Univ. Soc. Sci. 2014, 35, 115–122. [Google Scholar] [CrossRef]
  9. Lv, Z. Practice and Prospect of Expanding Social Participation in Cultural Heritage Conservation. Study Nat. Cult. Herit. 2025, 10, 1–3. [Google Scholar]
  10. Zhao, X.M. Cultural Rights in Complex Urban Heritage: Concept and Practice. Rev. Space Soc. 2024, 3, 113–123. [Google Scholar]
  11. Wu, W.Z.; Zhuang, Z.M. Design and Innovation of Tourism Products in the Era of the Experience Economy: A Case Study on the Experiential Development of Tourism Products in Ancient Villages. Tour. Trib. 2003, 18, 66–70. [Google Scholar]
  12. Chen, X.; Liu, X.R.; Ren, X.L. Research on Guest Experience Perception of Lingnan Rural Homestays Based on Online Review Texts. Public Commun. Sci. Technol. 2024, 16, 20–24. [Google Scholar] [CrossRef]
  13. Urry, J. The Tourist Gaze: Leisure and Travel in Contemporary Societies; Sage: London, UK, 1990. [Google Scholar]
  14. Ai, J.L.; Deng, Q.F. Traveling Experience Quality Appraisal Target System and the Assessment Method Grind. J. Hunan City Univ. 2008, 29, 51–54. [Google Scholar]
  15. Ma, T. Tourist Experience Quality and Satisfaction: Connotation, Relationship and Measurement. Tour. Trib. 2019, 34, 29–40. [Google Scholar]
  16. Stankov, U.; Gretzel, U. Tourism 4.0 technologies and tourist experiences: A human-centered design perspective. Inf. Technol. Tour. 2020, 22, 477–488. [Google Scholar] [CrossRef]
  17. Manning, C.D.; Raghavan, P.; Schütze, H. Introduction to Information Retrieval, 1st ed.; Cambridge University Press: Cambridge, UK, 2008; pp. 225–267. [Google Scholar]
  18. Hearst, M.A. Search User Interfaces; Cambridge University Press: Cambridge, UK, 2009; pp. 147–153. [Google Scholar]
  19. Riloff, E.; Shepherd, J. A Corpus-based Approach for Building Semantic Lexicons. Computation and Language. arXiv 1997. [Google Scholar] [CrossRef]
  20. Hatzivassiloglou, V.; McKeown, K. Predicting the semantic orientation of adjectives. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics and Eighth Conference of the European Chapter of the Association for Computational Linguistics, Madrid, Spain, 7–12 July 1997; pp. 174–181. [Google Scholar] [CrossRef]
  21. Alonso, M.A.; Vilares, D.; Gómez-Rodríguez, C.; Vilares, J. Sentiment Analysis for Fake News Detection. Electronics 2021, 10, 1348. [Google Scholar] [CrossRef]
  22. Li, D.N.; Liang, J. Text Mining of the “Shanghai Waste Sorting” Discourse on Sina Weibo: A Public Opinion Analysis Based on Python’s SnowNLP. Southeast Commun. 2019, 93–95. [Google Scholar] [CrossRef]
  23. Peng, M.; Xi, J.J.; Dai, X.Y.; He, Y.X. Collaborative Filtering Recommendation Based onSentiment Analysis and LDA Topie Model. J. Chin. Inf. Process. 2017, 31, 194–203. [Google Scholar]
  24. Karimi, K. The Configurational Structures of Social Spaces: Space Syntax and Urban Morphology in the Context of Analytical, Evidence-Based Design. Land 2023, 12, 2084. [Google Scholar] [CrossRef]
  25. Maniei, H.; Askarizad, R.; Pourzakarya, M.; Gruehn, D. The Influence of Urban Design Performance on Walkability in Cultural Heritage Sites of Isfahan, Iran. Land 2024, 13, 1523. [Google Scholar] [CrossRef]
  26. Ye, Y.; Zhang, L.Z.; Dou, W.T.; Zeng, W. Measuning Street Greening Quality from Humanistic Perspective: A Large-scaleAnalysis Based on Baidu Street View lmages and Machine Learning Algorithms. Landsc. Archit. 2018, 25, 24–29. [Google Scholar] [CrossRef]
  27. Wang, Z.Y.; Zhuang, W.M. Research on Block Renewal Post Occupancy Evaluation Methods Based on Image DeepLearning: A Case Study of Shichahai in Beijing. New Archit. 2022, 3, 5–8. [Google Scholar]
  28. Wang, J. Street Space Quality Evaluation from the Perspective of Urban Function. Master’s Thesis, Shandong Normal University, Jinan, China, 2021. [Google Scholar]
  29. Huang, J.X.; Liang, J.Q.; Yang, M.S.; Li, Y. Street Space Visual Quality Evaluation Method of Tourism Sites Based on Street View Images. J. Geo-Inf. Sci. 2024, 26, 352–366. [Google Scholar]
  30. Dai, Z.M.; Hua, C. The Improvement of Street Space Quality Measurement Method Based on Streetscape. Planners 2019, 35, 57–63. [Google Scholar]
  31. Zheng, Y.; Yang, J.Y. Research on Elaborate Urban Repair Planning Approach Based on AI Analysis of Large-scale Street-view Big Data. Chin. Landsc. Archit. 2020, 36, 73–77. [Google Scholar]
  32. Gu, W.Q. Study on Evaluation and Optimization Strategy of Street Spatial Quality in Sanyuan County. Master’s Thesis, Xi’an University of Architecture and Technology, Xi’an, China, 2024. [Google Scholar]
  33. Fang, Z.G.; Liu, C.; Xiao, Y.; Jin, C.H.; Zhuang, X.J. Evaluation of Street Aesthetics and Influencing Factors Using Multi-source Deep Learning Methods: The Case Study of Shanghai. Urban Plan. Int. 2023, 38, 48–58. [Google Scholar]
  34. Han, J.W.; Dong, L. A Study of Visual Evaluation of Streetscape Based on the Psychophysical Method. Chin. Landsc. Archit. 2015, 31, 116–119. [Google Scholar]
  35. Jinan Municipal People’s Government. Regulations on the Protection of Famous Historical and Cultural Cities in Jinan; Jinan Municipal People’s Government: Jinan, China, 2021. [Google Scholar]
  36. Luo, W.J.; Gu, H.Y.; Tao, W. Progress of space syntax research in both domestic and international geography. World Reg. Stud. 2024, 1–16. [Google Scholar]
  37. Qiao, W.Y.; Guan, W.H.; Wang, X.G.; Wang, X.; Gu, Z.L. Analyzing the Spatial Polycentric Evolution of Changsha City Based on the Space Syntax. J. Urban Reg. Plan. 2018, 10, 45–59. [Google Scholar]
  38. Zhang, L.; Ding, Y.D.; Zhang, X.Y.; Zhao, D.F. Research on Spatial Optimization Strategy of Historical Block Based on Spatial Syntax——Taking Zhongshu Street in Xuzhou as an Example. Urban Archit. Space 2022, 29, 32–35. [Google Scholar]
  39. Ji, W.M.; Ding, C.; Peng, B.Y. Analyzing the Spatial Form of Ancient Streets from the Syntactic Perspective: A Case Study of Kongcheng Old Street in Anhui. J. Xi’an Univ. Archit. Technol. Soc. Sci. Ed. 2024, 43, 54–61. [Google Scholar]
  40. Zhuang, L.H. Spatial Resilience Assessment and Enhancement Strategies for Residential Historic Districts: A Space Syntax-Based Approach. Master’s Thesis, Shandong Jianzhu University, Jinan, China, 2022. [Google Scholar]
  41. Ji, Z.M.; Tian, Z.G. Assessing the Accessibility of Qingfeng Park in Changzhou through Space Syntax Analysis. Agric. Technol. 2024, 44, 134–137. [Google Scholar]
  42. Qiu, J.; Huang, J.J. Research on optimization of urban recreation space based on space syntax. Shanxi Archit. 2024, 50, 24–27. [Google Scholar]
  43. Wang, C.; Yang, L.; Feng, X.; Li, Y.Y.; Zhen, X.H.; Weng, X.X.; Chen, S.J. Optimizing the Spatial Configuration of Industrial Historic-Cultural Districts: A Space Syntax-Based Strategy. Contemp. Hortic. 2024, 47, 108–110+114. [Google Scholar]
  44. Peponis, J.; Bafna, S.; Zhang, Z.Y. The connectivity of streets: Reach and directional distance. Environ. Plan. B Plan. Des. 2008, 35, 881–901. [Google Scholar] [CrossRef]
  45. Dou, Q. A Space Syntax Investigation of the Changes in Residential District Planning and Design in Beijing. Archit. J. 2010, 1, 28–32. [Google Scholar]
  46. Wang, L.H.; Zhang, C.N.; Xie, T. Research on Spatial Structure and Optimization of Chaoshan Village Based on Spatial Syntax and Social Network: Taking Chengyanggang Village in Shantou City as an Example. Ind. Constr. 2024, 54, 163–171. [Google Scholar] [CrossRef]
  47. Sun, Z.Y.; Han, B.R.; Wang, B.L. Research on Urban Three-dimensional Road Networks Based on Spatial Syntax: A Case Study of Nanjing Drum Tower. Logist. Sci.-Tech. 2024, 47, 71–74+78. [Google Scholar]
  48. Zhao, H.X.; Wen, J.; Gao, C.Y. Optimizing the Pedestrian Network of Daming Palace Heritage Park: A Space Syntax-Based Approach. Urban Archit. Space 2024, 31, 68–69. [Google Scholar]
  49. Wu, Y.Q.; Guo, S.Q.; Wang, C.S. Perceiving Hebei Province’s International Destination Image: An Analysis of User-Generated Content through the Lens of Appraisal Theory. J. Hebei Tour. Coll. 2025, 30, 29–35. [Google Scholar]
  50. Li, S.Y. Investigating the Relationships Between Live Streaming Commerce Content, Online Reviews, and Consumer Purchase Intention. J. Commer. Econ. 2023, 42, 59–62. [Google Scholar]
  51. Du, S.Y.; Wang, X.P. Study on Overseas Tourists’Perceptions of Huangshan Mountain’s Tourism Image Based on Tourist-generated Contents. J. Huangshan Univ. 2023, 25, 32–38. [Google Scholar]
  52. Cai, L.B.; Wan, F.L. Measurement, Analysis, and Optimization of Rural Tourism Public Service Quality: A Tourist Perception Perspective. J. Shandong Agric. Univ. Soc. Sci. Ed. 2025, 27, 83–93. [Google Scholar]
  53. He, W.Q.; Wang, X.W.; Wang, J.Q. Analyzing Jinan’s Destination Image Perception: A Textual Analysis of Online Travel Blogs. China Circ. Econ. 2025, 48–51. [Google Scholar] [CrossRef]
  54. Hong, B.X.; Liu, P.P. Assessing Cultural Tourism Experience Quality in Quanzhou: A Textual Analysis of Online Content. Bus. Econ. 2025, 52–54+62. [Google Scholar] [CrossRef]
  55. Li, X.X. Web Text-Based Study on Image Perception of Red Tourism in Lu’an Dabie Mountains. J. Jiamusi Vocat. Coll. 2025, 41, 76–78. [Google Scholar]
  56. Luo, J.; Luo, P.S. Mapping Destination Image through Digital Narratives and IPA Modeling: A Case Study of Bosten Lake in Xinjiang. Trade Fair Econ. 2025, 46–49. [Google Scholar] [CrossRef]
  57. Li, X.T. Deconstructing Spatial Contextuality: Elemental Analysis of Mingfu City’s Urban Palimpsest in Jinan. Master’s Thesis, Shandong Jianzhu University, Jinan, China, 2022. [Google Scholar]
  58. Du, H.X. Research on Landscape Renewal Design of Jinan General Temple Historical District under the Perspective of Residential Symbiosis. Master’s Thesis, Xi’an University of Architecture and Technology, Xi’an, China, 2023. [Google Scholar]
  59. Zhao, X.Q. An Analysis of the Conservation and Renewal of Urban Historical and Cultural Blocks From the Perspective of Cultural and Tourism Integration: A Case Study of Baihuazhou District, Mingfu City. Master’s Thesis, Guilin University of Technology, Guilin, China, 2022. [Google Scholar]
  60. Gu, F.Y. Research on the Relationship between Corporate Attention to Environmental Policy, Pressure from the Media, and Corporate Environmental Performance. Master’s Thesis, Nanjing Forestry University, Nanjing, China, 2023. [Google Scholar]
  61. Nan, J.M. Analysis of User Comments on Shared Study Rooms in China Based on Text Mining. Master’s Thesis, North Minzu University, Yinchuan, China, 2022. [Google Scholar]
  62. Zheng, M.Y. A Study on Sentiment Annotation Based on Functional Linguistics: A Case of Cross-Border E-Commerce Customer Reviews Between China and ASEAN. Ph.D. Thesis, Guangxi Minzu University, Nanning, China, 2024. [Google Scholar]
  63. Bird, S.; Klein, E.; Loper, E. Natural Language Processing with Python; O’Reilly Media: Sebastopol, CA, USA, 2009; pp. 121–135. [Google Scholar]
  64. Church, K.W.; Hanks, P. Word association norms, mutual information, and lexicography. Comput. Linguist. 1990, 16, 22–29. [Google Scholar]
  65. Zhang, S. Introduction to Conservation of Historic Cities: A Holistic Approach to Cultural Heritage and Historic Environment Protection (Third Edition). J. Tongji Univ. Soc. Sci. Sect. 2021, 32, 2. [Google Scholar]
  66. Li, H.P.; Wang, Y.H. Research on the Protection and Development of Historic Districts under the Measurement of Community Participation: A Case Study of the Ciqikou Historic District in Chongqing. China Anc. City 2024, 38, 3–9. [Google Scholar] [CrossRef]
  67. Wang, Y.; Xiu, C.L. Spatial Quality Evaluation of Historical Blocks Based on Street View Image Data: A Case Study of the Fangcheng District. Buildings 2023, 13, 1612. [Google Scholar] [CrossRef]
  68. Jiang, S.Y.; Liu, J. Comparative Study of Cultural Landscape Perception in Historic Districts from the Perspectives of Tourists and Residents. Land 2024, 13, 353. [Google Scholar] [CrossRef]
  69. Field, A. Discovering Statistics Using SPSS, 5th ed.; Sage Publications: London, UK, 2018. [Google Scholar]
  70. Stevens, J.P. Applied Multivariate Statistics for the Social Sciences, 6th ed.; Routledge: New York, NY, USA, 2020; pp. 329–347. [Google Scholar]
  71. Pei, Y.; Kan, C.; Dang, A. Street Greenspace Justice Assessment Study of Dongcheng District in Beijing Based on Street View Data. Chin. Landsc. Archit. 2020, 36, 51–56. [Google Scholar]
  72. Yang, J.Y.; Wu, H.; Zeng, Y. Research on Characteristics and Interactive Mechanism of Street Walkability Through Multi-source Big Data: Nanjing Central District as a Case Study. Urban Plan. Int. 2019, 34, 33–42. [Google Scholar] [CrossRef]
  73. Xiong, X.; Wen, X.Y.; Yang, S.R.; Liu, L.; Zeng, W. Visual Quality Assessment and Influencing Factors Analysis of Urban Waterfront Greenway Landscape: A Case Study of Nanjing Jiajiang Riverside Greenway. Landscape 2024, 41, 69–77. [Google Scholar]
  74. He, Z.Y. Study on the Personalized Design and Smart Management of the Advertising Signs on Urban Building Facade. Urban Archit. 2024, 21, 133–137. [Google Scholar] [CrossRef]
  75. Zhang, Y.C. Research on Quantification of Street Spatial Characteristics and Optimization Design in Zhengzhou Based on Street View Images. Master’s Thesis, Henan Agricultural University, Zhengzhou, China, 2024. [Google Scholar]
  76. Feng, X.X. Assessment and Optimization of Vegetation Shading Benefits Based on the Distribution of Human Flow in Campus. Master’s Thesis, Huazhong Agricultural University, Wuhan, China, 2024. [Google Scholar]
  77. Fang, H.Y. Research on Design Strategy of Pedestrian Space in Harbin Residential District from the Perspective of Vitality Creation. Master’s Thesis, Harbin Institute of Technology, Harbin, China, 2021. [Google Scholar]
  78. Jiang, L. Consuming the Urban Cultural Space through the Body: A Focus on Physical Bookstore Space. J. Shenzhen Univ. Humanit. Soc. Sci. 2023, 40, 57–66. [Google Scholar]
  79. Yin, S.M.; Guo, X.; Jiang, J. Extracting the Urban Landscape Features of the Historic District from Street View Images Based on Deep Learning: A Case Study in the Beijing Core Area. ISPRS Int. J. Geo-Inf. 2022, 11, 326. [Google Scholar] [CrossRef]
  80. Myers, J.L.; Well, A.D.; Lorch, R.F. Research Design and Statistical Analysis, 3rd ed.; Routledge: New York, NY, USA, 2010; pp. 521–525. [Google Scholar]
  81. Gibbons, J.D.; Chakraborti, S. Nonparametric Statistical Inference, 6th ed.; CRC Press: Boca Raton, FL, USA, 2020. [Google Scholar]
  82. Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Routledge: New York, NY, USA, 1988; pp. 20–27. [Google Scholar]
  83. Old Meets High-Tech! Fuzhou’s Three Lanes and Seven Alleys Fun Level up. Available online: https://mp.weixin.qq.com/s/cpUk3c0MXCe9CciQn6Hr9g (accessed on 10 June 2025).
  84. Creative AR Interactive Experience for Three Lanes and Seven Alleys Historical and Cultural District. Available online: https://www.imetacloud.com/plan/san-fang-qi-xiang-li-shi-wen-hua-jie-qu (accessed on 10 June 2025).
Figure 1. (a) China regional map. (b) Shandong regional map. (c) Jinan area map. (d) Jinan historical blocks distribution map. (Image source for (d): http://sd.ifeng.com/a/20161217/5240155_6.shtml, accessed on 25 April 2025).
Figure 1. (a) China regional map. (b) Shandong regional map. (c) Jinan area map. (d) Jinan historical blocks distribution map. (Image source for (d): http://sd.ifeng.com/a/20161217/5240155_6.shtml, accessed on 25 April 2025).
Buildings 15 02298 g001
Figure 2. (a) Historical resource distribution; (b) street network map of the study area.
Figure 2. (a) Historical resource distribution; (b) street network map of the study area.
Buildings 15 02298 g002
Figure 3. Research framework diagram.
Figure 3. Research framework diagram.
Buildings 15 02298 g003
Figure 4. Global integration value and local integration value.
Figure 4. Global integration value and local integration value.
Buildings 15 02298 g004
Figure 5. Connectivity value.
Figure 5. Connectivity value.
Buildings 15 02298 g005
Figure 6. Overlay diagram of integration value and connectivity value.
Figure 6. Overlay diagram of integration value and connectivity value.
Buildings 15 02298 g006
Figure 7. (a) Sampling point distribution; (b) street view image sampling.
Figure 7. (a) Sampling point distribution; (b) street view image sampling.
Buildings 15 02298 g007
Figure 8. Street view imagery crawling flowchart.
Figure 8. Street view imagery crawling flowchart.
Buildings 15 02298 g008
Figure 9. Unified image segmentation framework architecture. (Image Source: Adapted from the literature [79] “Extracting the Urban Landscape Features of the Historic District from Street View Images Based on Deep Learning: A Case Study in the Beijing Core Area”).
Figure 9. Unified image segmentation framework architecture. (Image Source: Adapted from the literature [79] “Extracting the Urban Landscape Features of the Historic District from Street View Images Based on Deep Learning: A Case Study in the Beijing Core Area”).
Buildings 15 02298 g009
Figure 10. Flowchart for calculating indicator proportions via semantic segmentation.
Figure 10. Flowchart for calculating indicator proportions via semantic segmentation.
Buildings 15 02298 g010
Figure 11. Scatter plot of street-level distributions for ten indicators.
Figure 11. Scatter plot of street-level distributions for ten indicators.
Buildings 15 02298 g011
Figure 12. Box plots: Different indicator scores across streets.
Figure 12. Box plots: Different indicator scores across streets.
Buildings 15 02298 g012
Figure 13. Box plots: Three-dimensional scores across streets.
Figure 13. Box plots: Three-dimensional scores across streets.
Buildings 15 02298 g013
Figure 14. Line chart: Expert and tourist ratings in three dimensions across streets. (a) Visual perception score; (b) Spatial comfort score; (c) Cultural experience score.
Figure 14. Line chart: Expert and tourist ratings in three dimensions across streets. (a) Visual perception score; (b) Spatial comfort score; (c) Cultural experience score.
Buildings 15 02298 g014
Figure 15. Heat-map: Indicator correlations in manual and machine scoring.
Figure 15. Heat-map: Indicator correlations in manual and machine scoring.
Buildings 15 02298 g015
Figure 16. Green view improvement schematic for Shuang Zhongci Street.
Figure 16. Green view improvement schematic for Shuang Zhongci Street.
Buildings 15 02298 g016
Figure 17. Facility layout adjustment plan for Hou Zaimen Street.
Figure 17. Facility layout adjustment plan for Hou Zaimen Street.
Buildings 15 02298 g017
Figure 18. Cultural scene integration schematic for Shuang Zhongci Street.
Figure 18. Cultural scene integration schematic for Shuang Zhongci Street.
Buildings 15 02298 g018
Table 1. Heritage Visibility Density of key streets.
Table 1. Heritage Visibility Density of key streets.
Street NameHeritage Visibility DensityStreet Length
Furong Street5.2 sites/100 mApprox. 420 m
Bai Huazhou Street4.5 sites/100 mApprox. 300 m
Jiang Junmiao Street3.8 sites/100 mApprox. 280 m
Hou Zaimen Street1.5 sites/100 mApprox. 200 m
Shuang Zhongci Street1.2 sites/100 mApprox. 150 m
Dong Huaqiangzi Street2.0 sites/100 mApprox. 250 m
Table 2. Composite Potential Index of streets.
Table 2. Composite Potential Index of streets.
StreetComprehensive Potential Index (CPI)InCnHVD (Sites/100 m)Typological Characteristics
Qu Shuiting Street (High-efficiency)2.0670.846 (High)0.561 (High)5.2 (High)Core cultural tourism corridors
Jiang Junmiao Street (Isolated)1.4110.5310 (Low)0.194 (Low)3.8 (High)Conflict between heritage conservation and accessibility
Shuang Zhongci Street (High-efficiency)0.8580.797 (High)0.598 (High)1.2 (Low)Local cultural deficiency
Hou Zaimen Street (Potential)0.7840.596 (High)0.318 (Low)1.5 (Low)Marginal revitalization blind spots
Shengfu Front Street (Potential)0.3270.789 (High)0.037 (Low)0Traffic-Induced deprivation of heritage space
Table 3. Sentiment statement filtering and scoring.
Table 3. Sentiment statement filtering and scoring.
Sentiment SegmentSentiment Score
Walking on Furong Street felt like being completely pushed along by the crowd—it was packed with people everywhere.−3
Furong Street is Jinan’s characteristic old street and food street, named after the famous spring “Furong Spring” on the street.3
Bordered by Mashi Street to the east, connecting to Qi Fengqiao Street, Xiangfeng Lane, and Furong Lane; adjacent to Yuhuanquan Street to the west, leading to Shengfu East Street.0
After admiring Qifeng Bridge, you can wander through the surrounding alleys to experience the charm of old Jinan.0
As a tourism educator and nature enthusiast, I personally believe that artificial attractions compromising natural landscapes are counterproductive.−1
The Qushuiting, Baihuazhou, and Mingfu City area south of the road represents Jinan’s spring water cultural characteristics and is the most worthwhile section to visit.1
Table 4. Word Frequency Statistics and Sentiment Scores of Physical Elements.
Table 4. Word Frequency Statistics and Sentiment Scores of Physical Elements.
Physical ElementWord FrequencySentiment Score
Spring Water8700, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 3, 3, 3, 3, 0, 0, 3, 3, 3, 3, 3, 0, 0, …
Lakeside6150, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, …
History4550, 0, −1, −1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 3, 3, 1, 0, 0, 1, 1, 1, 0, 0, …
Ancient Street2671, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1, 1, …
Decorated Archway151, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, −1
Balustrade30, 1, 1
Temple of Literature2390, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, −1, 0, −1, 1, 0, 0, 1, 1, …
Lao She2230, 0, 0, 0, 0, 1, 0, 3, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
Table 5. Core variables in principal component analysis (PCA).
Table 5. Core variables in principal component analysis (PCA).
Physical ElementTotal Word FrequencyMean Sentiment ScorePositive FrequencyNegative FrequencySentiment Variance
Spring Water8700.36196300.7933199771055678
Characteristic6371.4245940.9097815458958181
Lakeside6150.28143180.6894500687623522
Park5820.28125390.7774605564306287
Lotus5200.24112120.578430722378848
History4550.31117140.657227382560551
Culture4050.3911460.7095662231009212
Tourism4030.36117120.7068955872188105
Quancheng Square3920.2368150.660073757758151
Ancient Street2670.5612060.74885764957086
Table 6. Principal component factor scores.
Table 6. Principal component factor scores.
Visual Perception ElementsFactor 1Spatial Comfort ElementsFactor 2Cultural Experience ElementsFactor 3
Spring Water5.03821Lively Atmosphere2.27207Signboards2.88051
Park3.99731Trail2.12973Huixianlou Restaurant2.05298
Lakeside3.30701Hutong2.00042Listen to Crosstalk1.66994
Lotus2.52627Stalls1.85029Ancient Trees1.65663
Architecture0.78775Bustling1.75446Former Residence of Lu Dahuang1.65663
Pearl Spring0.62592Affordable1.6487Folk Performance1.65663
Photography0.88237Ancient Street1.23603Market1.65663
Flowers and Plants−0.5027Stone Bench−1.28178Historical Site1.65663
Blue Sky−0.48927Bench−1.2777Cultural and Artistic1.19729
Trees0.44884Shade of Trees−1.2777Former Residence of Lao She1.03973
Stream0.44837Leisure1.02609Storefront0.99678
Pond−0.40417Stroll1.00449Teahouse0.95303
Open Space−0.39282Quiet0.97863Lotus Festival0.84724
Advertisement−0.51878Decorated Archway0.96219Archway−1.29944
Street Lamps−0.49488Sculpture−1.27906Qifeng Bridge−0.81796
Table 9. Effect size of machine-derived metrics on human ratings.
Table 9. Effect size of machine-derived metrics on human ratings.
Machine IndicatorDimensionsMean of Compliant GroupMean of Non-Compliant GroupMean
Difference
Effect Size (Cohen’s d)
Green Vision RateVisual Score2.040.20+1.841.28 (Large Effect)
Shading Facility Coverage Visual Score2.040.20+1.841.28 (Large Effect)
Effective Pedestrian Clear WidthVisual Score1.850.31+1.540.99 (Large Effect)
Street Enclosure RatioVisual Score−0.381.60−1.98−1.44 (Large Negative Effect)
Green Vision RateComfort Score1.500.49+1.010.69 (Moderate Effect)
Effective Pedestrian Clear WidthComfort Score1.540.47+1.070.70 (Moderate Effect)
Shading Facility CoverageComfort Score1.580.44+1.130.77 (Moderate Effect)
Street Enclosure RatioComfort Score0.041.33−1.29−0.91 (Large Negative Effect)
Table 10. Mean proportion of key impact indicators versus mean three-dimensional ratings along streets.
Table 10. Mean proportion of key impact indicators versus mean three-dimensional ratings along streets.
StreetGreen Vision RateShading Facility CoverageEffective
Pedestrian Clear Width
Street Enclosure RatioVisual ScoreComfort ScoreCultural Score
Qu Shuiting Street0.2860.2560.0960.3142.5331.6940.858
Jiang Junmiao Street0.0760.0700.0520.5280.1210.6390.479
Shuang Zhongci Street0.0510.0470.0560.5430.0490.4100.578
Hou Zaimen Street0.0730.0620.0470.5560.6310.4850.970
Shengfu Front Street0.1920.1340.0620.3741.6341.5040.534
Table 11. Core deficiencies and strengths of streetscapes.
Table 11. Core deficiencies and strengths of streetscapes.
StreetCore DeficienciesStrengths
Jiang Junmiao Streetlow green vision rate (0.076), extremely low visual score (0.121)relatively high street enclosure ratio (0.528)
Shuang Zhongci Streetlowest green vision rate (0.051), lowest shading facility coverage (0.047), lowest comfort score (0.41)highest street enclosure ratio (0.543)
Hou Zaimen Streetlowest effective pedestrian clear width (0.047), low shading facility coverage (0.064), low comfort score (0.485)highest cultural score (0.97)
Shengfu Front Streetrelatively low shading facility coverage (0.134), low effective pedestrian clear width (0.062)acceptable visual/comfort score
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Chen, L.; Cai, X.; Liu, Z. Multi-Source Data and Semantic Segmentation: Spatial Quality Assessment and Enhancement Strategies for Jinan Mingfu City from a Tourist Perception Perspective. Buildings 2025, 15, 2298. https://doi.org/10.3390/buildings15132298

AMA Style

Chen L, Cai X, Liu Z. Multi-Source Data and Semantic Segmentation: Spatial Quality Assessment and Enhancement Strategies for Jinan Mingfu City from a Tourist Perception Perspective. Buildings. 2025; 15(13):2298. https://doi.org/10.3390/buildings15132298

Chicago/Turabian Style

Chen, Lin, Xiaoyu Cai, and Zhe Liu. 2025. "Multi-Source Data and Semantic Segmentation: Spatial Quality Assessment and Enhancement Strategies for Jinan Mingfu City from a Tourist Perception Perspective" Buildings 15, no. 13: 2298. https://doi.org/10.3390/buildings15132298

APA Style

Chen, L., Cai, X., & Liu, Z. (2025). Multi-Source Data and Semantic Segmentation: Spatial Quality Assessment and Enhancement Strategies for Jinan Mingfu City from a Tourist Perception Perspective. Buildings, 15(13), 2298. https://doi.org/10.3390/buildings15132298

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop