Sustainable Development Strategies for Culture–Tourism Integration in the Historic District of Tianzifang, Shanghai

Abstract

This study focuses on tourist-oriented urban historic districts. In recent years, many such districts have experienced commercial intensification and homogenization, placing pressure on sustainable development. The prior work is largely descriptive and offers limited mechanism-level guidance for governance. In response, this study employs Tianzifang as an empirical case and proposes an online-review-driven mechanism-identification framework. Drawing on 3005 online reviews, a quantitative–qualitative mixed approach was adopted: word-frequency and semantic-network analyses of the full corpus mapped topics and their relational structure; guided by these structures, grounded-theory coding was conducted on a negative-review subsample (n = 602); the results indicate a double-helix interaction between culture–commerce and expectation–reality, associated with lower perceived authenticity, affective disconnect, stronger negative word-of-mouth, and perceived declines in attractiveness. The main contributions are: a mechanism identification framework with a replicable quantitative–qualitative integration workflow; the construction of a double-helix mechanism coupling culture–commerce and expectation–reality; and, on this basis, a governance strategy framework to support fine-grained management and the sustainable renewal of urban historic districts.

1. Introduction

As carriers of local memory and community life, urban historic districts have long stood at the core of debates on cultural identity, urban regeneration, and cultural governance. UNESCO’s Recommendation on the Historic Urban Landscape (HUL) emphasizes that historic urban landscapes encompass not only individual buildings but also community life, local memory, and the multi-layered values of natural and cultural heritage, thereby constituting a vital resource for urban sustainability and cultural continuity [1]. In the experience-economy era and under culture–tourism integration, historic districts operate beyond cultural display: they are leveraged for city branding and the (re)construction of place identity [2,3], and serve as experience-led platforms that organize and activate tourism consumption spaces [4]. However, excessive commercialization has also led to spatial homogenization, staged and superficialized cultural expressions, and the marginalization of community life, triggering ongoing debates about the trade-offs between commercial gains and cultural values [5,6]. Against this backdrop, a demand-side perspective is essential. Visitors are the immediate experiencers of commercialization, and their visitation and consumption behaviors can feed back to reconfigure business formats and spatial structures [7]. Meanwhile, these visitor experiences are externalized via user-generated content (UGC), shaping destination imagery and influencing expectations for subsequent redevelopment [8,9]. From the visitor’s standpoint, perceived authenticity functions as a key criterion for judging the tension between commercialization and place-based values and has been linked to engagement, satisfaction, revisit intention, and word of mouth [10,11]. Beyond these outcomes, recent evidence shows that visitors’ willingness to pay more (WTP) is responsive to perceived benefits and is moderated by cultural value orientations [12]. Accordingly, this study adopts perceived authenticity as its core analytical entry point.

In recent years, research on tourism authenticity has advanced along two principal strands: (1) conceptual definition and typological clarification [13,14]; and (2) the operationalization and measurement of authenticity, followed by tests of its relationships with visitor outcomes and boundary conditions [15,16]. By contrast, the process through which visitors form authenticity judgments remains under-explained; methodologically, single-track designs still predominate and systematic mixed-methods studies are scarce [17], making triangulation between process evidence and statistical evidence difficult and thereby limiting the robustness of conclusions. To address these gaps and to strengthen comparability with international studies, this study—guided by the holistic and people-centered orientation of the UNESCO Recommendation on the Historic Urban Landscape (HUL)—develops a process–context analytical framework and undertakes a systematic examination of the generative mechanisms of visitor-perceived authenticity. To ensure methodological consistency and replicability, a single-source mixed-methods design is adopted, integrating quantitative techniques with qualitative narrative analysis on the same UGC corpus, with cross-validation at the interpretation stage.

To operationalize and test the framework, Shanghai is taken as the study setting and Tianzifang is selected as the focal case. As one of China’s most internationalized and market-oriented cities, Shanghai’s historic districts have long been situated in an intense culture–commerce interplay and continuous renewal, providing a typical context for examining the process–context framework. Within this setting, Tianzifang represents a lilong-type historic district that retains the shikumen urban fabric and everyday community life while exhibiting high visitor flows and market-oriented operations, making it suitable for identifying mechanisms under the culture–commerce tension. In parallel, the large volume of UGC accumulated on social media and online travel platforms provides a stable corpus for both quantitative and qualitative analyses. On this basis, an analytical framework for highly commercialized historic districts is proposed and tested, providing an actionable lens for observation and evaluation and, in turn, informing governance strategies with practical implementation recommendations.

2. Literature Review

2.1. Evolution of Historic District Conservation Concepts

Over the past century, the conceptual framework for conserving historic districts has undergone profound change. In the early twentieth century, heritage protection focused primarily on individual monuments, emphasizing the restoration of historical significance, architectural style, and structural form. This phase was object-centered and expert-led, with limited attention to the relationships between buildings and their surrounding environment or social functions [18,19]. As urbanization accelerated, many historic districts were demolished during redevelopment, provoking concerns about ruptures in place memory. In response, the Athens Charter (1931) argued that heritage protection should not be confined to isolated monuments [20]; the Venice Charter (1964) further underscored the inseparability of monuments and their settings and, through principles such as respect for original work and materials and the safeguarding of historical evidence, laid an early foundation for authenticity assessment in conservation practice [21]. By the late twentieth century, conservation thought expanded toward cultural diversity and the role of communities. The Amsterdam Declaration (1975) systematically proposed integrated conservation, stressing the incorporation of heritage into broader urban–rural development and social policy [22]; the Nairobi Recommendation (1976) advocated addressing the relationship between conservation and contemporary life at the scale of historic areas [23]; and the Washington Charter (1987) articulated principles for conserving historic towns and urban areas, emphasizing their linkage with community life [24]. This turn broadened the scope of conservation and introduced plural pathways for value assessment. With wider participation by diverse cultures, social groups, and institutions, disagreements intensified around what constitutes authenticity and who holds the authority to define it. Building on the Venice Charter’s emphasis on material and form, the Nara Document on Authenticity (1994) insisted that authenticity should be determined within specific cultural contexts and collective understandings, rejecting a single “universal” standard and expanding evidential bases to include function, traditional craftsmanship, setting, and even “spirit and feeling” [25]. Overall, conservation of historic districts has shifted from a material-based, expert-driven model toward collaborative governance that foregrounds cultural context and multi-stakeholder participation, thereby bringing public/visitor experience into view and providing a conceptual interface for subsequent research on authenticity and visitor perception.

2.2. Tourism Authenticity: Conceptual Evolution and Research Pathways

In parallel with the shift in historic–district conservation from “material–form” to “culture–context”, tourism studies have extended authenticity from an object attribute to a visitor-side issue of perception and experience, focusing on how authenticity is perceived, generated, and negotiated. The concept entered tourism research with Boorstin’s (1964) critique of “pseudo-events”, which argued that media and markets manufacture “consumable realities”, heightening modern anxieties about authenticity [26]. Building on this problem framing, MacCannell (1973) [5], drawing on Goffman (1959) [27], proposed staged authenticity, using the frontstage/backstage distinction to bring authenticity into sociological analyses of tourist motivation and experience. Since then, authenticity has become a central topic in tourism research.

A sustained conceptual and genealogical debate has unfolded internationally: Culler (1981) explained, from a semiotic perspective, how authenticity often emerges as symbolic reality through signs and projections [28]; Cohen (1988) emphasized its social construction and proposed emergent/gradual authenticity [29]; Bruner (1994) demonstrated, via case studies, the negotiation and reproduction of authenticity [30]; and Selwyn (1996) distinguished hot/cold authenticity to capture, respectively, participation–emotion and knowledge–evidence orientations [31]. Synthesizing this literature, Wang (1999) [32] systematized the field by typologizing objective and constructive authenticity and introducing existential authenticity. He further classified authenticity by its generative path: objective and constructive authenticity are defined as object-related, whereas existential authenticity is activity-related, thereby distinguishing “the truth of the object” from “the truth of the experience”. These typologies and generative paths have provided stable conceptual anchors for subsequent work on visitor perception and behavior, enabling operationalization at the levels of measurement and mechanism. Definitions of these authenticity constructs are provided in

Table 1. Key studies on tourism authenticity.

Author	Year	Title	Main Focus	Methods	Key Findings	Innovations/Contributions
Daniel J. Boorstin [26]	1964	The Image: A Guide to Pseudo-Events in America	Introduces “pseudo-events” and “consumable reality”; critiques media- and market-made ‘reality’.	Critical cultural analysis/conceptual exposition	Mass media and market logics can shape and amplify spectacles that appear ‘real’, influencing expectations and generating authenticity anxieties.	Positions the media–authenticity problem within modern communication/consumption; background for authenticity in tourism.
Dean MacCannell [5]	1973	Staged Authenticity: Arrangements of Social Space in Tourist Settings	Proposes “staged authenticity”; frontstage/backstage as lens on display and being watched.	Theoretical exposition + case synthesis	Tourist space is organized as ‘frontstage’; visitors judge true/false via front/backstage cues.	Centers display/being-watched in authenticity analysis; establishes a sociological perspective.
Jonathan Culler [28]	1981	Semiotics of Tourism	From semiotics: ‘symbolic’ authenticity formed through markers and texts.	Semiotics/discourse analysis	Authenticity is often constructed and consumed via signs such as ‘authentic/local/original’.	Introduces semiotic tools; highlights the role of texts, markers, and regimes of looking.
Eric Cohen [29]	1988	Authenticity and Commoditization in Tourism	Explains social construction of authenticity; proposes emergent/gradual authenticity.	Conceptual analysis and case discussion	‘Inauthentic’ items can be re-authorized as ‘authentic’ over time; commoditization does not necessarily dissolve authenticity.	Shifts from static ontology to an evolutionary/process perspective.
Edward M. Bruner [30]	1994	Abraham Lincoln as Authentic Reproduction: A Critique of Postmodernism	Shows authenticity as negotiated and reproduced among multiple actors.	Case study/interpretive anthropology	Authenticity is rewritten via narratives, power, and interests rather than being inherent.	Empirical demonstration of negotiated authenticity using a canonical case.
Tom Selwyn [31]	1996	Introduction (in The Tourist Image: Myths and Myth Making in Tourism)	Distinguishes ‘hot/cold’ authenticity (participation and emotion vs. knowledge and evidence).	Theoretical review/conceptual typology	Different orientations yield different authenticity judgments and experiential structures.	‘Hot/cold’ dichotomy bridges object-related and activity-related pathways.
Ning Wang [32]	1999	Rethinking Authenticity in Tourism Experience	Systematically reviews objective/constructive authenticity; introduces existential authenticity; proposes object-related vs. activity-related split.	Conceptual synthesis and theoretical analysis	Distinguishes ‘truth of the object’ vs. ‘truth of the experience’; objective/constructive = object-related; existential = activity-related.	Widely adopted ‘tripartite + dual-path’ framework enabling quantitative and contextual research.

.

Under this analytical framework, subsequent work has largely unfolded along two strands: (1) continued debates on definitions and types of authenticity [13,14]; and (2) a shift toward operationalization and measurement, followed by tests of authenticity’s effects on satisfaction, loyalty, and related outcomes, as well as the identification of pathways and boundary conditions [15,16]. Conceptually, divergences remain: authenticity is variously treated as an object property, a projection of attributes, or a mode of being [32,33]; other studies elaborate dimensions such as traditional representation, the “genuineness” of craftsmanship, negotiation processes, social imaginaries, and connections with the past [34]. Acknowledging this plurality, the present study adopts and pragmatically integrates a dynamic orientation [35,36], understanding authenticity as a judgment continuously generated and recalibrated through person–place interaction within specific situations. Accordingly, a process- and context-based perspective is adopted. Perceived authenticity is conceptualized as a relational judgment produced by the interplay between place cues and subjective expectations and interpretations, rendering the construct dynamic, complex, and context-dependent. On this basis, three testable threads are articulated: (1) prior-on-site consistency and its deviations; (2) the coupling between perceived authenticity and sense of place; and (3) differentiation of criteria and pathways induced by group heterogeneity.

In sum, authenticity resists reduction to a single dimension; it is a processual judgment co-shaped by mediated expectations (platform narratives, visual scripts) [37,38], on-site experience (built environment, everyday practices, place memory) [10,39], and population heterogeneity (cultural background, cohort, consumption orientation) [40]. Consequently, “appearing authentic/inauthentic” is dynamic and context-dependent, governed by multiple standards: prior expectations set evaluative coordinates; in situ cues and interactions provide or weaken experiential support; and diverse stances and preferences, within a communicative milieu of “being watched/being staged”, steer judgments and generate competing narratives. Put differently, destination authenticity is an ongoing negotiation whose fluctuations reflect alignments and misalignments among media scripts, place cues, and subjective expectations, intertwined with visitors’ affective identification and behavioral intentions. On this basis, the study employs an integrated process–context–mechanism framework to connect general patterns with context-sensitive explanations.

2.3. Methodological Review of Tourists’ Perceived Destination Image

Research on tourists’ perceived destination image generally adopts quantitative and qualitative approaches. Quantitative studies typically rely on structured scales to elicit evaluations of destination image attributes and apply statistical techniques to derive assessment results, which are often used in causal, explanatory, and predictive research [41,42]. Common methods include importance–performance analysis (IPA) [43], structural equation modeling (SEM) [44], and cluster analysis (CA) [45]. For example, Lee et al. (2023) used IPA to compare importance–performance gaps among international students in Korea [43]; Jebbouri et al. (2022) employed SEM to verify the pathway from image formation to satisfaction and trust [44]; and Karayazı et al. (2024) applied latent class analysis (LCA)—a model-based clustering technique—to segment tourists by heritage-site location preferences, identifying three groups and discussing their differentiated orientations toward urban heritage spaces and images [45]. Qualitative research, by contrast, often uses textual analysis (including online reviews, travelogs, and social media posts) and in-depth interviews, frequently in combination with grounded theory, to examine how tourists perceive destination image [46]. For instance, Ding (2024) analyzes the case of Meishan, China, using qualitative content analysis to map projected online destination images in the context of “all-for-one tourism” [47]; Küster et al. (2024) conduct netnography to study experience-oriented destinations, inductively developing qualitative themes from online discourse and linking them to concepts such as destination attachment, emotions, and loyalty [48].

Overall, quantitative research—grounded in structured measurement and statistical inference—offers strong objectivity and comparability, but is relatively limited in capturing complex contexts and generative mechanisms and in fully representing dynamic processes; qualitative research provides thick description of context and meaning and can capture nuanced emotions and narratives, yet its generalizability and replicability are comparatively constrained by researcher judgment and sampling. With respect to publication types, based on a systematic review and bibliometric analysis of tourism authenticity research covering 1979–2020 (tourism journals sample, N = 458), studies within that corpus still predominantly employ single-method designs—either quantitative or qualitative—while mixed-methods designs remain relatively scarce [17]. Guided by the research questions and data characteristics, this study adopts a mixed-methods design: the quantitative component first delineates overall structures, after which indicative cues are used to conduct grounded textual analysis that clarifies construct meanings and explains their contextual underpinnings. Evidence from the two strands is then triangulated at the stages of research design and interpretation, thereby leveraging their respective strengths, reducing single-path bias, and achieving quantitative–qualitative integration on the same corpus—a useful complement to the current literature in which mixed-methods studies remain relatively limited.

2.4. Review Summary

Historic-district conservation has transitioned from an expert-led, material–form paradigm to culture-context–based value judgments negotiated among multiple stakeholders. In tourism studies, debate has developed along the line of staged authenticity, followed by the codification of the objective–constructive–existential triad and the dual object- versus activity-based pathways; accordingly, the analytical lens has moved from object/symbol toward an experience dimension anchored in activity/self. Conceptually, divergences persist: authenticity is variously treated as an object property, a projection of attributes, or a mode of being; other work elaborates dimensions such as traditional representation, the genuineness of craftsmanship, negotiation processes, social imaginaries, and connections with the past. Rather than imposing a single definition, the present study turns to generation processes, understanding authenticity as a relational judgment produced in specific situations through the interaction of media scripts, place cues, and subjective expectations. Recent research has advanced in both breadth and depth, accumulating evidence on projection–reception consistency, the coupling of perceived authenticity and sense of place, and heterogeneity across groups. Nevertheless, most studies still proceed along a single facet and lack an integrated explanation within a unified framework. Methodologically, tourism-authenticity research remains dominated by single-method (quantitative or qualitative) designs, with mixed methods relatively scarce; in terms of data, one-off questionnaire/scale self-reports prevail, and systematic use of user-generated content (UGC) for authenticity remains limited.

To address these gaps, UGC from mainstream online platforms is employed as the core corpus, and media scripts–place cues–subjective expectations are incorporated into a single analytical framework. A linked mixed-methods design is implemented on the same corpus: the quantitative component delineates overall structures and key associations, and the qualitative component conducts semantic and narrative analysis to provide context-sensitive explanations of authenticity mechanisms, reducing “method juxtaposition” and single-path bias. To enhance comparability and auditability, operational definitions and coding principles are specified for core constructs. The technical pipeline and implementation details are presented in the Methods section, and empirical results appear in the subsequent sections.

3. Materials and Methods

3.1. Study Area

This study selects Tianzifang, located in Huangpu District, Shanghai, as the study area (see Figure 1). Formerly known as “Zhichengfang”, the area was renamed “Tianzifang” in 2002 by the contemporary artist Huang Yongyu. The new name is a homophonic reference to Tian Zifang, a legendary figure regarded as the founder of Chinese painting, symbolizing a place where artistic talents gather. Tianzifang is situated at the heart of the Dapuqiao commercial zone in central Shanghai, adjacent to the Sun Moon Light Shopping Center. It covers a total area of 7.2 hectares, with a 2-hectare core zone consisting of the “three lanes and one block”—Lanes 210, 248, and 274 off Taikang Road, as well as the street-facing section of Taikang Road [49]. Tianzifang is among the first districts in Shanghai recognized as a creative industry cluster and is widely regarded as a model for regional transformation driven by culture and the arts [50]. By introducing new commercial formats such as artist studios, designer shops, and specialty restaurants, the district has successfully transitioned from a traditional residential lilong (alleyway) neighborhood to a multifunctional urban space integrating cultural tourism, commercial consumption, and local community life. This study selects Tianzifang as the research subject based on the following considerations.

Firstly, Tianzifang retains the shikumen–lilong spatial fabric to a high degree of integrity, making it one of the few lane-and-alley settlements in central Shanghai that still preserve authentic character; its urban grain continues that of the early twentieth century. Meanwhile, the district presents a juxtaposition of Chinese traditional residential typologies and Western stylistic details—combining vernacular dwellings and shikumen/new-style lilong with classical, modernist, and even Baroque elements—thus embodying the Sino-Western hybrid evolution characteristic of old Shanghai’s historic districts [51].

Secondly, through an integrated model of “historic architecture + creative experience + tourism consumption”, Tianzifang maintains the early 20th-century shikumen esthetic while introducing a wide range of functions such as galleries, designer boutiques, cultural and creative markets, bars, and traditional street food [52]. This fusion creates a spatial atmosphere that combines deep historical and cultural heritage with vibrant modern creativity. Home to hundreds of creative commercial and cultural spaces, the district has become one of Shanghai’s most popular tourist landmarks.

Thirdly, Tianzifang has accumulated a vast volume of tourist review data across various social media platforms. This provides a rich source of user-generated content for researchers to analyze tourists’ perceptions of the district, particularly regarding their multidimensional needs in cultural experience, spatial atmosphere, commercial formats, and service quality.

Fourth, Tianzifang faces several challenges common to historic districts. While the surge in tourist numbers has brought significant economic benefits, the growing popularity has also led to a range of negative effects that undermine the district’s long-term sustainability—ultimately resulting in reduced foot traffic and a decline in overall attractiveness. How to simultaneously preserve cultural heritage, manage commercialization, and improve the tourist experience has become a central problem that urgently needs to be addressed in Tianzifang’s future development.

Currently, scholarly research on Tianzifang primarily focuses on spatial production [53], cultural hybridity/integration [51], and spatial vitality and regeneration [54]. Zhong examines interactions among heterogeneous actors in spatial practices and their effects on urban spatial restructuring in Tianzifang [53]; Xu analyzes the multicultural hybridity of Tianzifang’s urban space and its causes [51]; and Shan et al. assess the vitality of the renewed historic built environment and discuss regeneration strategies for cases such as Tianzifang [54]. However, tourist-side perceptions of authenticity and the underlying psychological mechanisms remain underexplored. Only a few studies touch on visitor experiences and behavioral responses in Tianzifang—for example, analyses of tourist satisfaction in cultural-creative destinations [55] and the relationship between collective memory and travel intention from a tourist-gaze perspective [56]. Accordingly, this study aims to elucidate the generative pathways of visitor-perceived authenticity in Tianzifang and to assess its behavioral effects.

3.2. Research Framework

The research workflow is depicted in Figure 2. First, data-acquisition platforms consistent with the research objectives and data availability were selected. In compliance with platform terms and research ethics, Python 3.8.7 was used to scrape and aggregate tourist reviews, yielding an initial sample. Second, the raw texts underwent preprocessing—including content screening, synonym consolidation, correction of typos/misspellings, and deduplication—to construct a valid review corpus. The standardized corpus was then imported into ROST-CM6 for parallel text analytics: (1) high-frequency word analysis to identify salient elements and overall attentional foci; (2) semantic network analysis based on high-frequency terms to construct the co-occurrence network and identify core structures and key associative pathways; and (3) sentiment analysis to determine polarity (positive/neutral/negative), which was used to delineate a negative-review subset as the sampling frame for the qualitative stage. Next, the negative subset was imported into NVivo 15 for open, axial, and selective coding, enabling mechanism discovery and inductive conceptual model building. Finally, outputs from the three computational procedures were integrated and triangulated with qualitative evidence to produce a governance strategy framework for historic districts.

3.3. Data Sources and Processing

To ensure the authority and representativeness of the dataset, this study selects platforms from major domestic travel-booking portals and local life-review sites. The selection follows three criteria: (1) platform coverage and user base; (2) standardization of object scope and data fields (i.e., whether reviews are aggregated at the “attraction/district” level and whether core fields such as review text, star rating, and timestamp are provided); and (3) data accessibility and review volume. According to recent monitoring statistics [57], Ctrip maintains a leading market share (approximately 35.04%), followed by Qunar (15.9%), Meituan (14.3%), and Tongcheng Travel (11%). In terms of monthly active users (MAU), Ctrip recorded about 71.7 million MAU in 2021, ranking first in the industry, followed by China Railway 12306, Tongcheng Travel, Meituan, and Qunar. To cover both “destination sightseeing evaluations” and “in-district consumption and experience”, in addition to travel-booking/destination-review platforms, Dianping is included as the representative local life platform. A preliminary screening using “Tianzifang” as the keyword was then conducted across candidate platforms to compare review counts, update frequency, and content completeness. The pretest results indicate that Dianping and Ctrip perform better on these indicators and are suitable for subsequent analysis. Accordingly, Dianping and Ctrip are finalized as the data sources for analyzing tourists’ perceived image of Tianzifang.

Considering the impact of COVID-19 and the importance of data timeliness, the data collection period was set from 8 January 2023, to 31 May 2025. On 8 January 2023, China’s National Health Commission officially announced the removal of mandatory centralized quarantine and nucleic acid testing for inbound travelers, and downgraded COVID-19 from a Class A to a Class B infectious disease, signaling a return to normalcy in domestic tourism activities [58]. Using Python, the researchers scraped relevant reviews from both Dianping and Ctrip, initially collecting 3121 review entries. The data were then carefully filtered and processed according to the following steps: ① Content Filtering: Reviews were excluded if they lacked substantive content (e.g., simple copies of Tianzifang’s introduction, image-only uploads, advertisements), contained too few words, were incomplete or repetitive, or were only marginally related to tourist experiences. This ensured the practicality and quality of the data. ② Synonym Replacement: A synonym dictionary was developed to standardize semantically similar terms. For example, “shop” and “small shop” were unified as “shop”; similarly, “foreigner”, and “foreign tourist” were all unified as “foreigner”, ensuring consistency and accuracy in statistical analysis. ③ Typos and Misspelling Corrections: Common errors in tourist reviews, such as misspellings of “Tianzifang” and “shikumen”, were corrected to ensure lexical consistency. After this filtering process, a total of 3005 valid review entries were retained for analysis (see

Table 2. Summary of online text data sources.

Platform	Original Sample Size	Valid Samples After Processing	URL
Dianping	3032 comments	2923	https://www.dianping.com (accessed on 30 June 2025)
Ctrip Travel	89 comments	82	https://flights.ctrip.com (accessed on 30 June 2025)
Total	3121	3005

).

3.4. Data Analysis Methods

3.4.1. Content Analysis

Content analysis originated in the field of communication studies and was defined by Bernard Berelson as “a research technique for the objective, systematic, and quantitative description of the manifest content of communication” [59]. It remains one of the most fundamental methods in communication research [60]. This method is widely used for quantitative analysis of texts or communicative content, including theme extraction, sentiment identification, and model construction. In recent years, content analysis has been extensively applied in studies on tourist perception, destination image, and related areas [61,62]. In this study, based on the filtered online reviews of Tianzifang, multi-level textual analysis was conducted using ROST Content Mining 6 (ROST-CM6) software. Developed by Professor Shen Yang of Wuhan University, ROST-CM6 is a specialized computational text analysis tool capable of performing word segmentation, Chinese-English word frequency statistics, semantic network analysis, and sentiment analysis. It is widely used in the processing and study of online textual data [63].

3.4.2. Grounded Theory

Following the identification of emotional tendencies in tourist reviews through content analysis, this study further employs grounded theory to conduct in-depth qualitative analysis of the negative reviews. The aim is to uncover the underlying factors contributing to negative tourist perceptions and, subsequently, propose strategies for the sustainable development of historic districts. Grounded theory was formally introduced by Glaser and Strauss in 1967 in their seminal work The Discovery of Grounded Theory, where they emphasized that theory should be “grounded” in empirical data rather than driven by pre-established hypotheses [64]. This methodology has had a profound influence on qualitative research in sociology, communication studies, management, and other disciplines.

In the field of tourism research, Hottola was among the first to apply grounded theory to the study of cross-cultural adaptation in tourism, achieving notable research outcomes [65]. In recent years, grounded theory has gained increasing traction in tourism studies, particularly in research on cross-cultural behavior, tourist experience, and adaptation. It has become a widely adopted approach for theory generation in qualitative tourism research [46]. Grounded theory can be categorized into three main schools: Classic Grounded Theory, Pragmatic Grounded Theory, and Constructivist Grounded Theory [66]. Among these, the Pragmatic Grounded Theory emphasizes a clearer and more structured process of data analysis. By standardizing and making the coding procedures transparent, it enhances the traceability and consistency of the research process. Therefore, this study adopts Pragmatic Grounded Theory as the methodological framework to ensure the systematic and scientific rigor of the analysis.

4. Data Analysis and Research Findings

4.1. Content Analysis

4.1.1. High-Frequency Word Analysis

The pre-processed textual data were saved as a TXT file and imported into ROST-CM6 for functional analysis. The text was first segmented into words, and then the “word frequency analysis” function was applied to generate a word frequency statistics file. The resulting data were manually filtered to remove meaningless or irrelevant high-frequency words, retaining only those that were semantically relevant to the research topic. After processing, the study identified the top 60 high-frequency words that reflect tourists’ perceived image of Tianzifang (see

Table 3. High-frequency words from online texts related to Tianzifang.

Rank	High-Frequency Words	Frequency	Part of Speech	Rank	High-Frequency Words	Frequency	Part of Speech	Rank	High-Frequency Words	Frequency	Part of Speech
1	Tianzifang	2426	noun	21	Lively	251	adjective	41	History	145	noun
2	Shanghai	1923	noun	22	Flavor	250	noun	42	Retro	142	adjective
3	Shop	1176	noun	23	Tourism	250	noun	43	Studio	141	noun
4	Lane	977	noun	24	Food	246	noun	44	Charm	140	noun
5	Check-in	686	verb	25	Bar	240	noun	45	Transportation	136	noun
6	Specialty	659	noun	26	Culture	239	noun	46	Commercial	132	adjective
7	Stroll	583	verb	27	Friend	226	noun	47	Experience	131	verb
8	Not bad	467	adjective	28	Feel	209	verb	48	Interesting	128	adjective
9	Shikumen	409	noun	29	District	198	noun	49	Style	128	noun
10	Tourist	377	noun	30	Evening	185	noun	50	Atmosphere	126	noun
11	Snacks	361	noun	31	Subway	177	noun	51	Environment	121	noun
12	Take photos	359	verb	32	Convenient	172	adjective	52	Coffee	120	noun
13	Foreigner	348	noun	33	Dapuqiao	171	noun	53	Modern	119	adjective
14	Art	327	noun	34	Popularity	166	noun	54	Delicious	115	adjective
15	Architecture	320	noun	35	Qipao	160	noun	55	Attract	105	verb
16	Riyueguang	296	noun	36	Pandemic	155	noun	56	Walk	105	verb
17	Attraction	292	noun	37	Worth it	153	verb	57	Vanishing Cream	104	noun
18	Alley	281	noun	38	Artistic	151	adjective	58	Souvenir	103	noun
19	Cuisine	270	noun	39	Weekend	147	noun	59	Handmade	101	noun
20	Creativity	257	noun	40	Taikang Road	145	noun	60	Accessories	101	noun

).

From a part-of-speech perspective, the extracted high-frequency words were predominantly nouns, totaling 43 (72%), mainly related to the destination’s location, environment, food, architecture, and transportation. In addition to place names like “Tianzifang” and “Shanghai”, frequently occurring terms included “store”, “lane”, and “shikumen”, which together sketch an overall image of the district as perceived by tourists. There were 9 adjectives (15%), most of which expressed positive or neutral evaluations, such as “not bad” and “interesting”, indicating that tourists generally held favorable views of Tianzifang, though some opinions were reserved. Notably, the appearance of the term “commercialization” reflects concern among some tourists regarding the district’s level of commercial development. Finally, 8 verbs (13%) were identified, mostly relating to typical tourist activities in Tianzifang, such as “stroll”, “take photos”, “check in”, “experience”, and “feel”, illustrating behavioral characteristics centered on sightseeing, social engagement, and light cultural exploration.

Based on word-frequency statistics and rank ordering, the high-frequency terms in Table 3 were classified by semantic orientation to more intuitively portray the “culture-commerce” tension in online perceptions. Terms associated with historical morphology and local practices were labeled “culture-oriented” (e.g., shikumen, longtang [lanes], history, retro, literary, culture, art, qipao), whereas terms related to consumption and attention/traffic attributes were labeled “commerce-oriented” (e.g., shops, check-ins, photo-taking, delicacies, bars, souvenirs). Neutral/functional terms (e.g., Shanghai, metro, transportation, attraction, Taikang Road, Dapuqiao) were excluded from the comparison. Among the top 20 terms, five were commerce-oriented—shops (1176; rank 3), check-ins (686; rank 5), snacks (361; rank 11), photo-taking (359; rank 12), and food (270; rank 19)—with a cumulative frequency of 2852. Four were culture-oriented—longtang (977; rank 4), shikumen (409; rank 9), art (327; rank 14), and creativity (257; rank 20)—with a cumulative frequency of 1970. Commerce-oriented terms cluster around the “check-in–photo-taking–snacks/food–shops” chain of visitation and consumption activities; culture-oriented terms center on “longtang–shikumen–art/creativity,” linking lane morphology with esthetic/creative elements. Collectively, the online discourse simultaneously points to specific actions within consumption settings and to the spatial/cultural attributes of the historic quarter, yielding a direct quantitative depiction of the “culture–commerce” tension under examination.

4.1.2. Semantic Network Analysis of High-Frequency Words

While high-frequency word analysis can reveal patterns in word usage and frequency distribution within the text, it often falls short in uncovering the semantic relationships between terms, and does not capture the structural patterns or associative networks among them. To address this limitation, this study employed the “Semantic Network Analysis of High-Frequency Words” function provided by ROST CM6 software to process the high-frequency terms in the Tianzifang reviews and construct a semantic network graph (see Figure 3). In this network, nodes represent individual words, with larger nodes indicating higher word frequency; edges (connections) between nodes represent word co-occurrence relationships, with thicker lines denoting stronger associations. Semantic network analysis belongs to the domain of network science and, through graph theory techniques, allows for the visualization of the semantic structure among words. It serves as an effective tool for uncovering patterns of cognitive association and understanding [67].

As shown in Figure 3, the semantic network of Tianzifang presents a typical “core–periphery” structure. The core skeleton is composed of the nodes “Tianzifang–Shanghai–shops–lanes,” supplemented by high-centrality terms such as “check-in”, together forming the main semantic framework of tourist reviews. The skeleton indicates that tourists tend to situate Tianzifang within the broader urban imagery of “Shanghai,” linking commercial consumption experiences, represented by “shops”, with the traditional spatial form symbolized by “lanes”. This connection not only reflects the close integration of commercial activities and the spatial environment of historic districts but also reveals that lane culture in contemporary tourism narratives is increasingly reconstructed and reproduced within a commercialized context. The “check-in” node maintains medium-strength connections with the core skeleton, suggesting that its role as a means of social sharing and visit verification is dependent on the main chain of “Tianzifang–Shanghai–shops” rather than forming an independent thematic center. In contrast, terms such as “culture”, “art”, and “creativity” appear on the periphery of the network and exhibit relatively weak linkages with the core nodes, indicating their lower salience in tourist reviews and their tendency to appear as supplementary descriptors rather than as central elements in shaping the perception framework. Overall, the structure demonstrates that tourists’ perceptions of Tianzifang are primarily shaped by the interplay of urban context, commercial experiences, and traditional spatial imagery, while cultural and creative elements play a comparatively limited role in shaping actual perceptions.

4.1.3. Sentiment Analysis

Using the “Sentiment Analysis” function in ROST-CM6 software, tourist reviews were statistically analyzed (see

Table 4. Sentiment analysis of tourist reviews.

Sentiment Category	Size	Proportion (%)
Positive Sentiment	2358	78.47
Neutral Sentiment	45	1.50
Negative Sentiment	602	20.03
Total	3005	100

). The results show that positive sentiment accounts for 78.47% of the texts, negative sentiment for 20.03%, and neutral sentiment for 1.50%. Overall, the online image of Tianzifang is predominantly positive, which aligns with the earlier high-frequency word analysis that revealed a majority of favorable terms. However, approximately one-fifth of the reviews contained negative sentiments, indicating the presence of dissatisfaction or disappointment in tourists’ actual experiences. Based on this finding, further content analysis was conducted on the 602 negative reviews to explore the sources of dissatisfaction and the underlying cognitive mechanisms.

4.2. Grounded Theory Analysis

To further investigate the negative sentiments identified through sentiment analysis, this study employed grounded theory methodology, using NVivo 15 software to conduct open coding, axial coding, and selective coding on each of the 602 negative reviews. The aim is to explore the reasons behind tourists’ negative perceptions during their experiences in Tianzifang and to reveal the internal mechanisms driving these responses.

4.2.1. Open Coding

Open coding involves comparing, organizing, summarizing, and coding the existing data in order to conceptualize and categorize it. In this phase, conceptual clustering of the original data yielded 21 initial concepts. The detailed data structure for the open coding results is presented in

Table 5. Open coding results.

Category (Node Count)	Conceptualization	Sample Quotes from Original Text
A1 Product Homogenization (376)	Lack of distinctive products; an overabundance of national tourism commodities	“They’re selling the same stuff you see everywhere—no different from Chenghuang Temple or Nanluoguxiang”. “All the snack and souvenir shops are identical, completely lacking uniqueness”.
A2 Overpricing (342)	Price fraud/tourist scams; nostalgic products severely overpriced	“Coconut went from ¥15 to ¥35—pure scam”. “Childhood nostalgia items that used to cost cents now go for over ten yuan”.
A3 Low-End Commercial Leasing (208)	Night market-style leasing; declining commercial quality	“Now filled with fried tornado potato shops, neon signs blinding your eyes”. “Feels like a flea market with fake book covers”.
A4 Wave of Store Closures (287)	Large-scale vacancy; transfers; significant decline in commercial vitality	“Over a third of the shops are closed—transfer signs everywhere”. “On a Saturday night, the whole alley is pitch dark”.
A5 Sharp Decline in Foot Traffic (401)	Tourism carrying capacity left unutilized; strong sense of desolation	“Electronic screen shows 5000 max capacity, but only 200 people inside”. “More staff than tourists—depressingly empty”.
A6 Deteriorating Sanitation (264)	Poor management of public hygiene; conflict between modern and traditional habits	“The alleys stink, and garbage is everywhere”. “A resident dumped a chamber pot in the street—it’s 2024”!
A7 Dilapidated Infrastructure (221)	Decay and disrepair; lack of environmental maintenance	“Potholes filled with water; wires exposed like a spiderweb”. “Walls haven’t been painted in ten years—plaster falling off”.
A8 Filthy Toilets (187)	Severe lack of public restrooms; unsanitary conditions	“Only one squat toilet with a 30 min wait—so filthy it’s unusable”. “The toilet collapsed—it’s a disgrace to Shanghai”.
A9 Spatial Congestion (172)	Overcrowded physical spaces; conflict between residential and tourist spaces; safety risks	“Barely enough room for two people to walk side-by-side—and underwear hanging overhead”. “Stalls are blocking emergency exits”.
A10 Parking Difficulties (98)	Insufficient parking; poor management; expensive fees	“Security guards just shoo cars away—parking here is hell”. “¥15/hour at Riyueguang hurts the wallet”.
A11 Lack of Wayfinding (67)	Dysfunctional signage; chaotic route design	“Google Maps led me around for 20 min—I couldn’t find the shop”. “Seven entrances feel like a maze”.
A12 Loss of Uniqueness (369)	Devaluation of place spirit; loss of cultural atmosphere; authenticity undermined; tourism trap	“The soul of Shikumen is drowned in neon—now it’s a monstrosity”. “The original residents are gone—only tourist traps remain”.
A13 Decline of Art (214)	Shrinking art spaces/activities; commercialization and tokenization of art	“Chen Yifei gallery is shut—only skincare products being sold”. “All the artists left—art is now just a backdrop”.
A14 Loss of Historical Atmosphere (192)	Destruction of historic appearance; discordant mixture of old and new	“Historic buildings are covered by shop renovations—like a beggar covered in jewelry”. “Laundry poles and neon lights make for a grotesque mix”.
A15 Resident Conflicts (123)	Invasion of privacy; tension between residents and tourists	“Got yelled at by a resident—‘No photos allowed!’” “Underwear flapping above tourists’ heads”.
A16 Ineffective Marketing (84)	Long-term absence of promotion; lack of storytelling	“Haven’t seen any promotional activity in three years”. “They don’t even know how to tell a good story anymore”.
A17 Negative Recommendations (318)	Strong word-of-mouth dissuasion; local identity crisis	“Absolutely don’t come—even locals feel embarrassed”. “Not coming is a one-hour regret; coming is a full-day regret”.
A18 Poor Service (276)	Forced purchases; verbal abuse; poor attitude	“Bead shop forced add-ons—cussed at me when I refused”. “Tried on a qipao—got chased and yelled at for not buying”.
A19 Food Traps (243)	Poor quality and overpriced food; forced transactions	“¥30 for fermented soup that’s supposed to be dessert”. “Charged ¥25 for fried sausage and forced me to pay”.
A20 Monotonous Experience (195)	Poor experiential quality; lack of content; overemphasis on photo ops	“Takes 10 min to walk through—not even as big as the mall restroom”. “Other than taking pictures, there’s nothing to do”.
A21 Tourist Scams (157)	Price manipulation; deceptive sales; hidden charges	“Fruit tea listed as ¥1.2/gram but charged ¥600/kilogram”. “So-called free internet-famous shop ended up charging ¥300”.

.

4.2.2. Axial Coding

Axial coding refers to the process of clustering and integrating open codes based on the logical relationships among initial categories, thereby forming broader and more abstract categories. Following axial-coding logic and assisted by NVivo 15, the initial 21 subcategories were grouped into higher-order categories. Through an integrative analysis, six core categories were ultimately identified: Commercial Degradation, Environmental Disorder, Cultural Dilution, Management Incompetence, Experiential Breakdown, and Tourist Resistance. The detailed axial coding structure is presented in

Table 6. Axial coding structure.

Core Category	Subcategory	Subcategory Description
C1 Commercial Degradation	B1 Product Devaluation	Severe product homogenization; lack of local features; low product quality and deceptive practices eroding tourist trust.
	B2 Price Disorder	Excessive and non-transparent pricing; price fraud; nostalgic goods severely overpriced, offering poor value for money.
	B3 Commercial Decline	Widespread shop closures; low-end vendor influx; shift from artistic to low-tier food services; declining vitality, high rent pressure, and frequent tenant turnover.
C2 Environmental Disorder	B4 Declining Foot Traffic	Sharp drop in visitor numbers; weekday/night-time desolation; more staff than tourists; loss of “internet-famous” appeal.
	B5 Infrastructure Decay	Aging infrastructure; scarce and filthy restrooms; blocked fire exits pose safety risks.
	B6 Traffic Barriers	Inadequate parking, high fees; aggressive vehicle management; confusing circulation design and ineffective navigation.
	B7 Sanitation Collapse	Accumulated garbage and pervasive odors in public spaces; clash between residential and commercial hygiene practices.
C3 Cultural Dilution	B8 Loss of Distinctiveness	Authenticity of Shikumen architecture eroded; transformed into a superficial tourist trap; local cultural symbols over-commercialized and devoid of depth.
	B9 Artistic Withdrawal	Shrinking art spaces and artist exodus; art reduced to background décor; creative industries replaced by generic vendors.
	B10 Resident Conflict	Spatial contestation between residents and tourists; original residents displaced due to commercialization, leading to the collapse of indigenous community culture.
C4 Management Incompetence	B11 Planning Vacuum	Chaotic spatial layout and undefined branding; homogenized commercial leasing; lack of mechanisms to balance cultural preservation and commercial needs.
	B12 Maintenance Failure	Deteriorated buildings left unrepaired; exposed, tangled electrical wires; lack of sanitation supervision and delayed waste removal.
	B13 Marketing Stagnation	Long-term absence of promotion; failure in storytelling; reliance on outdated travel guides with no updates aligned to market changes.
C5 Experiential Breakdown	B14 Service Violations	Forced purchases, verbal abuse; price deception and coercive sales practices.
	B15 Resource Scarcity	Insufficient rest areas; lack of diverse and engaging experiences beyond photo ops; poor signage and wayfinding difficulties.
	B16 Health Risks	Food safety concerns; cramped spaces increasing respiratory infection risk; exposed wiring and blocked fire exits posing major hazards.
C6 Tourist Resistance	B17 Behavioral Resistance	Active discouragement of others from visiting; refusal to spend money; preference for alternative destinations.
	B18 Emotional Aversion	Strong feelings of disappointment; crisis of local identity; anger toward over-commercialization.

.

4.2.3. Selective Coding

Selective coding refers to the process of further comparing and analyzing the core categories and their interrelations as established in the axial coding phase, in order to identify a core category that integrates all other categories into a coherent theoretical framework. By describing the structural relationships among the key categories, selective coding helps to analyze how the core category interacts with other major themes, and ultimately leads to the development of a new qualitative model. Based on the results of axial coding, it becomes clear that the six major categories—Commercial Degradation, Environmental Disorder, Cultural Dilution, Management Incompetence, Experiential Breakdown, and Tourist Resistance—form a well-defined causal chain. Among them, Management Incompetence emerges as the primary driving force behind tourists’ rising negative perceptions of Tianzifang. It directly contributes to both commercial degradation and environmental disorder, while also enabling the process of cultural dilution. Meanwhile, commercial degradation and environmental disorder further accelerate and intensify cultural dilution, which in turn causes experiential dissonance, eventually triggering tourist resistance. Tourist resistance then feeds back negatively into management performance, creating a vicious cycle. The interaction pathways between the major categories and the core category are illustrated in

Table 7. Selective coding structure.

Typical Relationship Structure	Interpretation of Relationship
Management Incompetence → Commercial Degradation	Absence of planning and lack of maintenance result in loss of control over business formats: homogeneous leasing, price inflation, influx of low-end vendors; artistic spaces are replaced by low-quality commercial operations.
Management Incompetence → Environmental Disorder	Neglect of infrastructure maintenance and failed traffic planning lead to spatial disarray: restroom scarcity, exposed wiring, blocked fire exits, dysfunctional wayfinding systems.
Commercial Degradation → Cultural Dilution	Over-commercialization squeezes out cultural space: product homogenization weakens local distinctiveness; shrinking artistic sectors lead to “Shikumen’s soul being buried under neon lights”; original residents displaced by commercial pressures, resulting in the collapse of community culture.
Environmental Disorder → Experiential Breakdown	Spatial congestion and sanitary collapse directly undermine visitor experience: narrow alleys, foul smells, and safety risks evoke an urge to “escape within 10 min”; infrastructural deficits exacerbate health hazards.
Cultural Dilution → Experiential Breakdown	Loss of cultural authenticity causes emotional disconnection: the withdrawal of art and the vanishing of uniqueness create a sense of “disillusionment,” and the absence of historical ambiance weakens immersive experience.
Experiential Breakdown → Tourist Resistance	Service violations and resource scarcity provoke resistance: forced consumption and food scams lead to “strong dissuasion” behavior; emotional aversion results in a crisis of local identity.
Tourist Resistance → Management Incompetence	Negative word-of-mouth undermines governance effectiveness: declining visitor flow accelerates commercial decay; falling revenue limits investment in maintenance, forming a “decline–neglect–further decline” vicious cycle.
Core Double Helix Conflict Mechanisms	Culture–Commerce Spiral: Commercial expansion → Symbolization of cultural space → Decline in authenticity → Decrease in foot traffic → Further reliance on commercial exploitation—forming a self-reinforcing loop. Expectation–Reality Spiral: Marketing shapes high expectations → Experiential dissonance → Emotional rupture → Intensified marketing to offset bad reviews → Expectations rise even higher.

.

4.2.4. Saturation Test

To ensure the scientific rigor and conceptual completeness of the grounded theory analysis, this study strictly adhered to the procedures of Pragmatic Grounded Theory and employed a dual validation mechanism to assess theoretical saturation:

The first step involved determining the point of repetition during open coding. By the time the 487th negative review (accounting for 80.9% of the sample) was analyzed, no new initial concepts emerged. Frequently appearing concepts such as “product homogenization” (A1) and “artistic withdrawal” (A13) continued to recur with similar expressions. The remaining 115 reviews were analyzed for verification, and no new concepts or relational patterns were discovered, thus satisfying the saturation criterion of “repetition of concepts and absence of new information”.

A logical closure among the major categories had been achieved. The six core categories identified through axial coding—Commercial Degradation, Environmental Disorder, Cultural Dilution, Management Incompetence, Experiential Breakdown, and Tourist Resistance—fully encapsulate the initial subcategories, and their interaction pathways form a closed-loop structure. To test theoretical robustness, 30 extreme negative reviews were retroactively analyzed. In each case, the model successfully explained the source of conflict, such as: “Over-commercialization → Cultural disappointment → Emotional rupture → Resistance behavior”.

Through this dual verification—data-level repetition and theoretical closure—the category system and theoretical model developed in this study are confirmed to be saturated. The model can systematically explain the formation mechanisms of tourists’ negative perceptions in Tianzifang.

4.2.5. Theoretical Model Construction

Based on the three-level coding results of the negative reviews, this study integrates the findings to propose a “Double-Helix Conflict” Theoretical Model of tourist perception in Tianzifang historic district (see Figure 4). The model reveals the existence of two interwoven conflict chains that underpin tourist perceptions: Culture–Commerce Conflict Chain: The advance of commercialization weakens cultural space and diminishes perceptions of authenticity, which in turn leads to a decline in tourist appeal and reputation. Expectation–Reality Conflict Chain: The gap between high visitor expectations (often shaped by media and marketing) and the actual on-site experience causes a rupture in emotional identification. These two conflict chains interact dynamically in an evolving feedback loop, constituting the underlying mechanism behind the weakening attractiveness and increasing polarization of tourist perceptions in Tianzifang historic district.

5. Discussion

5.1. Summary of Main Findings

Drawing on online tourist reviews, a progressive sequence—high-frequency term analysis, semantic network analysis, sentiment analysis, and grounded theory—is employed to move from surface description to underlying mechanisms in depicting the structure of visitor perceptions and their affective consequences. Word-frequency results indicate that visitor narratives align along two concurrent strands: one is a visitation–consumption–social chain centered on “check-in–photograph–food–shops”, and the other is a spatial/historical imagery chain represented by “longtang–shikumen–art/creativity”. The former dominates at the high-frequency level, suggesting that Tianzifang is more often situated within a scenarized, consumable interpretive frame; the latter appears primarily as label-like nouns, with relatively fewer processual and affective expressions. Meanwhile, the emergence of “commercialization” signals attention to—and disagreement over—the intensity of development, producing a semantic imbalance characterized by stronger consumption orientation and the backgrounding of cultural meanings. The semantic-network analysis structurally corroborates and deepens these observations: the network exhibits a core–periphery configuration; narratives are organized along the backbone path “Tianzifang–Shanghai–shops–longtang”, with “check-in” functioning as an intermediary node attached to the main chain. “Culture/art/creativity” lies at the periphery and is weakly connected to the core, indicating that actual perceptions are shaped primarily by the assemblage of “urban context–commercial experience–traditional space”, while the organizing force of culture/creativity is relatively limited.

Sentiment analysis provides an aggregate quantification of evaluative polarity: overall positive. However, online word-of-mouth is not an equal-weight average of positive and negative comments; rather, it is governed by weighting effects. Weights depend on the strength of cues, their positions within the semantic network (i.e., whether they occur at key touchpoints along the main corridor), and their visibility. Negative experiences occurring at critical touchpoints possess higher diagnosticity and narrative specificity and, through platform interactions, tend to attract greater attention, often producing notable rating penalties with amplification. By contrast, positive statements are typically mild and dispersed instances of routine satisfaction, making them less able to offset strong negative cues. Consequently, even with an overall positive polarity, a small number of high-intensity negative reviews located at key touchpoints become cognitively “up-weighted”, exerting a disproportionate downward pull on the overall impression. Accordingly, grounded coding was conducted on negative reviews to identify high-diagnostic touchpoints and their pathways. On this basis, a Double-Helix Conflict Model was inductively constructed, comprising two interrelated chains—the culture–commerce conflict chain and the expectation–reality conflict chain—to systematically explain the drivers of divergent visitor perceptions and the process by which Tianzifang’s attractiveness as a historic district weakens. Subsequent sections elaborate the triggering conditions, transmission links, and amplification mechanisms within this framework.

5.2. Theoretical Mechanisms and Interpretation of the Double-Helix Conflict Model

5.2.1. The “Culture–Commerce” Spiral Conflict

Coding reveals that references to “artistic fading/character loss” (classified under cultural dilution) frequently co-occur with mentions of commercialization within the same review, forming an upstream evidentiary cluster for the culture–commerce chain. The process can be summarized as follows: symbolization, commodification, and labeling/branding reshape the field’s meaning structure; cultural production and in situ interaction are subsequently compressed (e.g., creative/exhibition spaces gradually give way to homogenized retail and services), thereby weakening the conditions for generating existential authenticity. In its early phase, Tianzifang—marked by artist residency and residents’ everyday life—saw cultural production and local interaction jointly support visitors’ sense of presence. As commercialization advanced, creative spaces shifted into sales outlets, exhibitions closed, and handcrafts and human interactions were diluted. Evaluative phrases such as “soulless”, “an empty shell”, and “doesn’t feel like this place” frequently appear alongside these changes, corroborating a chain-like evolution whereby rising commercialization leads to cultural repackaging and contraction of local interaction, which in turn diminishes existential authenticity.

Moreover, declining authenticity often propagates via polarized word-of-mouth and reduced attractiveness into operational decision-making, which tends to be answered with short-cycle commercial responses (e.g., intensified leasing and promotional campaigns). Such responses further raise the commercial share, squeeze cultural-production space again, and produce a self-reinforcing loop.

5.2.2. The “Expectation–Reality” Spiral Conflict

Before the visit, platform communications and social media construct an a priori image of Tianzifang as “artsy, niche, and creative”. On site, however, cues such as over-commercialization, product homogenization, crowding, and service mismatch diverge from that image, producing an expectation–disconfirmation gap. When actual experience falls below expectations, overall evaluations and satisfaction decrease markedly, accompanied by negative word-of-mouth and avoidance intentions.

In terms of authenticity types, the core mechanism operates at the level of constructive authenticity: the destination image built by media and platform scripts is disconfirmed or shown to be incongruent on site; the resulting negative affect and evaluations spill over to existential authenticity, lowering visitors’ sense of presence and self-congruence. Typical review phrasings—e.g., “came because of X/online said it’s artsy—on site it’s commercial”, “not worth it/regret/don’t come”—co-occur with cues such as “crowded, homogenized, long queues, poor service,” revealing a continuous process of “image incongruence → negative disconfirmation → negative evaluation”. Correspondingly, declines in word-of-mouth and attractiveness are often met by amplified promotion and supplemental marketing, which elevate pre-visit expectations in the next cycle and thus magnify subsequent disconfirmation.

5.3. Recommendations for the Sustainable Development of Tianzifang Historic District

In view of the underlying mechanisms of the double-helix conflict identified in tourist perceptions, Tianzifang should proactively transform these conflicts into synergies by taking “cultural value” as its foundation, “experiential innovation” as its engine, “digital empowerment” as its bridge, and “refined services” as its safeguard. The goal is to reverse the two spirals of conflict into a mutually reinforcing, upward spiral, thereby constructing an internally driven and self-sustaining mechanism for long-term development.

5.3.1. Reinforce the Spirit of Place to Drive Business Optimization and Spatial Revitalization

This strategy aims to preserve and reinforce the spirit of place embodied by shikumen architecture and the unique East–West architectural hybridity of Old Shanghai, transforming it into a source of commercial vitality rather than allowing it to be diluted by commercial forces. Specifically, it encompasses three dimensions: establishing and enforcing culturally oriented business entry and renewal standards, integrating cultural narratives into spatial revitalization, and fostering a “brand manager culture”. First, culturally oriented entry and renewal standards should be developed. Priority should be given to introducing and nurturing businesses that are deeply integrated with Shikumen culture, Shanghai style art, and creative design, while restricting the expansion of homogenized, low-end, and overly commercialized chain stores. Second, spatial renewal should serve cultural storytelling. Under the premise of preserving the historic architectural character, underutilized or idle spaces should undergo micro-renovations to enhance their cultural narrative and readability, enabling the architecture itself to become both an “exhibit” and an integral part of the visitor experience. Finally, a “brand manager culture” should be actively cultivated. Independent brand owners with cultural commitment and innovative capacity should be encouraged to settle in, creating a distinctive “shop owner culture” that adds personality, charm, and a sense of human warmth to the district.

5.3.2. Innovate Experience Models and Build Open Participation Platforms

This strategy can be implemented through thematic immersive interactive programs, open creative studios, community tours with cultural interpretation, and interactive installations with contribution walls. The aim is to break away from the passive sightseeing model, enabling visitors to shift from “spectators” to “participants” and “co-creators”, thereby enhancing the depth and satisfaction of their experiences while bridging the “expectation–reality” gap. Thematic immersive interactive programs may include culture-based workshops, artist-in-residence creation showcases, micro-theater street performances, and cultural treasure-hunt games, all designed to encourage deep visitor engagement. Open creative studios allow visitors to observe, and even take part in, parts of the creative process up close, enhancing the authenticity and uniqueness of the experience. Community tours with cultural interpretation involve inviting local residents, veteran shop owners, or cultural scholars to serve as guides, offering personalized tours with warmth and historical depth to tell the “Tianzifang story”. Interactive installations and contribution walls can be placed in public spaces to host activities such as message-leaving, drawing, or photo-taking, with immediate public display. These invite visitors to share thoughts, creative ideas, or artworks, forming a continuously updated “visitor memory wall”.

5.3.3. Leverage Digital Technology to Build a Resilient Emotional Community

Digital technologies can overcome the limitations of physical time and space, extend the district’s influence, and foster lasting emotional connections that go beyond a single visit—thus enhancing resilience against future risks. Firstly, a high-precision “digital twin” of Tianzifang should be created using technologies such as 3D scanning, VR/AR, and the Internet of Things (IoT). This virtual district would faithfully replicate its architectural features, unique shops, and ambient atmosphere online. Secondly, a “virtual–physical integrated” global tourist community should be developed based on the digital twin. Functions could include virtual tours, immersive shop showcases, livestream shopping, online workshops, and digital collectible releases. Establishing forums, interest groups, and a membership system would connect tourists worldwide who are interested in Tianzifang’s culture, creating a sustained online platform for interaction, sharing, and co-creation. Thirdly, online–offline linkage should be encouraged. Highlights from offline events and creative outputs can be showcased and disseminated through the online community, encouraging further reinterpretation. Conversely, high-quality content and popularity from the online platform can drive real-world foot traffic, forming a virtuous cycle of engagement.

5.3.4. Establish Agile Feedback Mechanisms to Enhance Precision Management and Services

Going forward, it is essential to establish an agile feedback mechanism to capture the pulse of visitor experiences in real time, respond promptly to needs and issues, and use human-centered, refined management and services to improve visitor satisfaction and positive emotions while resolving potential conflicts. A multi-dimensional visitor perception monitoring network can be deployed, including installing intelligent foot-traffic and emotion-analysis devices at key points; setting up convenient feedback channels both online and offline; and conducting regular in-depth interviews and focus group discussions. Additionally, a closed-loop “perception–analysis–response” system can be established. A dedicated team would be responsible for real-time collection, multi-dimensional analysis, and rapid assignment of feedback for handling. Standardized contingency plans and quick-response procedures should be in place for high-frequency issues and emergencies. Thirdly, human-centered service details should be promoted. This includes optimizing wayfinding signage systems; adding rest areas, shaded/rain-shelter spots, nursing rooms, and accessible pathways; improving public restroom cleanliness and maintenance frequency; strengthening security patrols and flexible crowd management; and providing training for merchants and service staff to enhance service awareness and communication skills. Finally, a periodic “Experience Optimization Report” can be published along with acknowledgments, showing the public the improvements made based on feedback, and expressing gratitude to visitors who provided constructive suggestions—demonstrating both respect and commitment.

5.4. Limitations, Boundary Conditions, and Directions for Future Research

This study systematically characterizes the structural features of visitor perceptions of Tianzifang, proposes a Double-Helix Conflict Model, and discusses its operative mechanisms and key influencing factors. Several limitations remain, pointing to directions for improvement.

(1)

Limits of data sources and discursive context. The corpus primarily comprises online reviews from Chinese-language platforms. Influenced by cultural context and platform affordances, the content tends to foreground spatial and consumption carriers while underrepresenting emotion and identity expression. Future research may expand samples within a multi-language, cross-platform design and employ bilingual/multilingual collaborative coding with back-translation to enhance cross-cultural generalizability and sample representativeness.

(2)

Lack of empirical model validation. The current analysis is grounded in text mining and theory. Category relations and the theoretical model have not yet been empirically validated using quantitative methods. Subsequent work is encouraged to apply structural equation modeling (SEM) or survey-based designs to test the Double-Helix Conflict mechanism and to examine the robustness of its causal pathways.

(3)

Single-case limitation. Focusing on a single case (Tianzifang) constrains external validity. Future studies can adopt comparative multiple-case designs encompassing historic quarters of different types and developmental stages to further assess the model’s applicability and adaptability.

(4)

Scope conditions and external validity. The proposed mechanisms are more likely to hold in historic quarters featuring fine-grained block morphologies (e.g., lane-based fabrics), incremental renewal, mixed business formats, and multi-actor collaborative governance, and in discursive/platform environments comparable to those examined here. In contexts characterized by large-scale area redevelopment, single-actor strong governance, marked differences in visitor composition, or substantial divergences in language and platform mechanisms, the magnitude, timing, and drivers of effects may differ. Direct transplantation is therefore inadvisable; verification across multiple cases should precede context-specific adjustments.

6. Conclusions

Using Tianzifang as the empirical field, this study adopts a single-source mixed-methods design on a UGC corpus, integrating quantitative text analytics with grounded-theory coding, and articulates a Double-Helix Conflict Model. The quantitative analyses—word-frequency profiling, co-occurrence analysis and semantic-network mapping, and centrality metrics—map the topic structure and salience pattern, indicating that visitor narratives are more sensitive to spatial and operational cues, whereas meaning-related cues (culture, art, on-site creation) are comparatively attenuated. Guided by this structural map, grounded coding provides a mechanism-oriented interpretation of high-salience and key co-occurrence touchpoints, yielding two intertwined pathways: a culture–commerce chain, in which commercial intensification is associated with compressed cultural visibility and diminished perceived authenticity; and an expectation–reality chain, in which gaps between media-amplified expectations and fragmented on-site experiences are commonly accompanied by affective disconnect. Coupled through platform mechanisms, the two chains form a representation–governance–reconfiguration feedback loop; together, the quantitative structure and qualitative mechanisms mutually corroborate a process- and context-based explanation for asynchronous shifts in popularity and reputation in historic districts.

Theoretically, the study integrates object-side commercialization–authenticity tensions and subject-side expectation–experience inconsistencies within a unified process–context framework; methodologically, it provides an integrated quantitative–qualitative workflow on a single-source UGC corpus; and practically, it outlines a process-oriented governance approach focused on coordinating operational timing, optimizing the use–space (program/tenant) mix, and aligning and strengthening place narratives.