Mapping Digital Solutions for Multi-Scale Built Environment Observation: A Cluster-Based Systematic Review

Highlights

What are the main findings?

• Four distinct methodological clusters were identified (Data Integration and Us-er-Centric Analysis, Advanced 3D Spatial Analysis and Processing, Real-Time Inter-action and Digital Twin Support, and 3D Visualization) that define the current digital solutions for built environment observation spanning multi-scale applications from object to urban and heritage levels.
• A sequential and interdependent workflow across these clusters was revealed, from integrated data inputs to communicative 3D representations, highlighting their com-plementary yet fragmented nature.

What is the implication of the main finding?

• Combining identified clusters into interoperable frameworks can transform frag-mented prototypes into holistic platforms for decision support and urban planning.
• Advancing the field requires shifting from isolated functionalities toward systemic, interoperable architectures that bridge data, analysis, interaction, and visualization.

Abstract

This study investigates the intersection of digital tools and methods with the built environment disciplinary framework, focusing on Urban Planning and Development (UPD), Architecture, Engineering, and Construction (AEC), and Cultural Heritage (CH) domains. Using a systematic literature review of 29 solution-oriented documents, the research applies both bibliometric and in-depth content analysis to identify methodological patterns. Co-occurrence mapping revealed four thematic clusters—Data Integration and User-Centric Analysis, Advanced 3D Spatial Analysis and Processing, Real-Time Interaction and Digital Twin Support, and 3D Visualization—each corresponding to distinct stages in a digital workflow, from data acquisition to interactive communication. Comparative and interdependency analyses demonstrated that these clusters operate in a sequential yet interconnected manner, with Data Integration forming the foundation for analysis, simulation, and visualization tasks. While current solutions are robust within individual stages, they remain fragmented, indicating a need for systemic interoperability. The findings underscore the opportunity to develop integrated digital platforms that synthesize these clusters, enabling more comprehensive observation, management, and planning of the built environment. Such integration could strengthen decision-making frameworks, enhance public participation, and advance sustainable, smart city development.

1. Introduction

1.1. General Background

In the constantly evolving landscape of urbanization and urban transformation, we are experiencing the era of twin transition as a spot where digital amplifies sustainable growth [1]. The notion of twin transition recognizes that there is a significant and predominantly unutilized potential for technology to propel sustainability goals, including ‘making cities and human settlements inclusive, safe, resilient, and sustainable’ (SDG 11) [2]. In this still emerging notion, the twining of twins opens the question of how this innovative strategy could strengthen different domains of professional activity related to (1) understanding the dynamics of development and transformation of the built environment, and (2) identifying areas for optimization which could result in more efficient use of land and resources. At the outset, this implies the need to systematically research and understand the digital advancements for observing the built environment, which are strongly integrated into the notions of Smart Cities [3,4,5] and City 4.0 [6]. Moreover, the advent of new geospatial data and tools [7] with their operational capacity contribute to the increasing illumination of physical spaces as an interweaving historical, socio-economic, and geographical context [8].

The technological progress shift has fundamentally reshaped numerous sectors, including the interdisciplinary hybridized field of the built environment [9]—standing out as a prime example of this evolution. Professionals and policy makers in Architecture, Engineering, and Construction (AEC) and Urban Planning and Development (UPD) domains engage the spectrum of digital advancements in order to ensure data-driven and evidence-based practices by integrating various technological advancements like the Internet-of-Things [10], artificial intelligence (AI) [11,12], and its subfields of machine learning (ML) [13] and deep learning [14,15], big data [16,17,18], digital twins [19,20,21], blockchain [22,23], extended reality (XR) technologies encompassing augmented, virtual, and mixed reality [24], geographic information systems (GIS) [25], building information modeling (BIM) [26], computer-aided design (CAD) [27], 3D modeling and visualization [28], urban simulation and modeling [29], and digital mapping and surveying [30], etc. The introduction of digital-based methodologies has not just updated conventional planning and design approaches but has also enhanced their efficiency and set the stage for a comprehensive digital overhaul for built environment observation.

Accordingly, this research is concentrated on the application of digital methodologies from a multiscale perspective—from cityscape and landscape, over neighborhoods and districts, to individual buildings and plots. Besides the spatial scale of the observation, the domain of Cultural Heritage (CH) significantly problematizes the temporal scale of the built environment, providing invaluable insights into the evolution, significance, and legacy of human settlements over time and from a morphogenetic perspective. In the past decade, various digital tools and methods for the built environment observation (including analysis, simulation, representation, and decision making and planning support) have been introduced and implemented. However, there is a general lack of procedures for comprehensive observation of the built environment concerning both different scales and activities during planning and design processes. Accordingly, this paper recognizes the importance of a synchronous observation of the multi-domain phenomenon of digital advancements—within AEC, UPD, and CH—with the intention to gain a comprehensive insight into the existing body of methodologies for observing the built environment.

1.2. Motivation: Ongoing Research Trends and Thematic Trajectories

In parallel with the technological and methodological advancements discussed above, digital transformation in the built environment is increasingly guided by a comprehensive framework of international and European policy, legislative, and standardization initiatives. At the global level, the New Urban Agenda (2016) [31] and the UN Sustainable Development Goals (2015) [2]—particularly SDG 11 and the SDG Digital Acceleration Agenda (2023) [32]—position digitalization as a key driver of sustainable, inclusive, and data-informed urbanization. The UN–Habitat People–Centered Smart Cities Program (2025) [33] and the United for Smart Sustainable Cities (U4SSC) initiative (2025) [34] further promote interoperable and human-centered approaches to smart city development. Within the European context, the European Green Deal (2019) [35], the New Leipzig Charter (2020) [36], and the EU Industrial Strategy—Construction Transition Pathway (2023) [37] explicitly identify digital technologies as enablers of the green and circular transition. Legislative instruments such as the Public Procurement Directive 2014/24/EU (2014) [38], the Energy Performance of Buildings Directive (EPBD -Revised) (2024) [39], and the Construction Products Regulation (2024) [40] institutionalize this shift through mechanisms including BIM, Digital Building Logbooks, and Digital Product Passports. Complementary initiatives—such as DigiPLACE (2020) [41], the European Data Space for Smart and Sustainable Cities and Communities (DS4SSCC) (2022) [42], and the EU Local Digital Twins Toolbox (2025) [43] advance operational implementation, while standardization efforts (ISO 19650 Series, CEN/TC 442 BIM) and the emerging European Framework for Building Digital Logbooks (2020) [44] ensure coherence across domains.

Against this policy-driven framework, there is a growing body of academic literature that investigates how digitalization reshapes the AEC, UPD, and CH domains. The following sub-sections provide an overview of these scholarly trajectories, identifying thematic and methodological trends that inform the present study.

1.2.1. Advancements in Urban Planning and Development (UPD) Domain

In the fast-evolving landscape of the UPD domain, digital advancements stimulated frameworks for predict-oriented methodologies to align future actions toward innovative solutions [45]. Studies involving monitoring and assessment of land use change through employment of GIS-based methods for measuring environmental impact [46] and mapping of cross-scale spatial indicators [47], as well as remote sensing techniques [48] demonstrated the capacity of digitally based approaches to simulate the urban future. Change detection and future prediction methods significantly contributed to the introduction of space-time exploratory approach [49] and more robust approaches to visualization and analysis of spatiotemporal patterns [50]. These advancements include (1) interactive mapping and geospatial analysis [51], (2) multimodal and multidimensional geodata visualization [52], (3) collaborative visual analysis [53], and (4) metrics and modeling through geo-informatics [54].

In strengthening the temporal and spatial dimensions in visualization, the need for advancements is also recognized in the necessity of strengthening semantic visualization, both at the general level of representation and in the 3D urban environment [55]. Visual representations of semantic information (such as land use classifications, building functions, or environmental characteristics) have the potential to enable better perception and understanding of additional layers of meaning and relationships among urban features and elements. Unlike general visualization, which primarily focuses on the visual display of spatial data, semantic visualization enriches the interpretative dimension and supports more informed analytical insights. Semantic visualization further initiated a completely new understanding of urban patterns, which, in addition to spatio-temporal patterns, also include behavioral patterns that refer to the ways in which users interact with and utilize the built environment [56,57,58]. In this framework, digitally based approaches have demonstrated significant capacity for mapping behavioral patterns, including integrating geovisual analytics with ML [59], but also advanced technologies in urban games and engagement [60,61,62], which could provide an exploratory approach for urban data visualization and spatial analysis.

1.2.2. Advancements in Architecture, Engineering, and Construction (AEC) Domain

One of the standout innovations driving the digital transformation of AEC domain is the adoption of Digital Twins and BIM [63]. The fusion of these technologies-initiated capabilities for visualization, simulation, and optimization of different aspects related to project’s lifecycle including (1) sensing and monitoring of built environment through semantic enrichment of BIM [64], (2) leveraging digital twin models for zero-energy districts [65], (3) construction asset management using BIM [66], and (4) standardized specifications [67]. Moreover, the assessment-based engagement of digital models to analyze climate-related risks [68] and conduct Life Cycle Sustainability Assessments [69] significantly contributes to the improvement of ecologically and resilience-oriented approaches for monitoring building comfort and safety [70], as well as for virtual sensing in intelligent buildings and digitalization processes [71].

XR technologies (Augmented (AR), Virtual (VR), and Mixed (MR) reality) have also emerged as novel drivers of design and construction methodologies, offering real-time feedback and digital interaction [72,73]. Moreover, XR applications and devices exemplify the AEC domain’s embrace of cutting-edge solutions to (1) present construction information in both mixed reality and virtual reality environments [74], (2) perform data analysis and assessment [66], (3) integrate diverse datasets into a unified interface for both on- and off-site usage [75], (4) visualize spatial building data (e.g., building systems energy consumption information) [76] or live environmental data (such as indoor air temperature, light intensity, and humidity) [77] or (5) monitor and document construction site progress [78]. In parallel, reality capture techniques [79] have impacted construction project management by enabling real-time insights and informed decision-making.

1.2.3. Advancements in Cultural Heritage (CH) Domain

Digital advancements have revolutionized the CH domain and transferred impact towards UPD and AEC domains, in terms of documentation, reconstruction, dissemination, and representation of heritage assets. Heritage building documentation and digital reconstruction have been significantly enhanced through terrestrial laser scanning and satellite data [80]. These techniques enable precise capture of architectural details on a building level scale, providing the foundation for accurate digital reconstructions [81]. Scan methods and tools are primarily applied in the context of the reconstruction of built environments, serving as the drivers for the development of digital twins that replicate heritage sites [82]. Also, multi-source image fusion technology [83] together with the digital methods for visual inspection and condition assessment [84] enhances digital reconstructions particularly in terms of structural integrity and conservation needs of heritage assets. In this context, evolution of BIM technology into Historic/Heritage BIM (HBIM) marks a significant advancement in the preservation and analysis of built heritage through employing it as a multifunctional tool for (1) operationalization of 3D heritage reconstruction [81], (2) heritage analysis [85], (3) heritage communication through virtual tours [86], and (4) algorithmic modeling of complex construction elements [87].

The heritage dissemination domain is enriched through a serious games and procedural modeling [88], while virtual reconstructions [89] in both 2D and 3D formats bring historic sites to realistic context, allowing users to explore and interact with cultural heritage through gamification [90]. Additionally, heritage representation is elevated through laser scanning and 3D visualizations [91], which enable cultural resilience and community engagement. Representation and visualization of urban fabric through historical documents [92] further enriches the narrative of cultural heritage, highlighting the evolution of cities and the significance of historic landmarks.

1.3. Paper Outline and Objectives

Relying on recognized research trends and thematic trajectories in the UPD, AEC, and CH domains in relation to digital advancements, a three-fold need is recognized: (1) to establish a comprehensive overview of methodological clusters shaping digital solutions in the built environment, (2) to decode the key features and attributes that define the functionalities of these solutions, and (3) to reflect on their systemic interplay and future applicability across contexts and scales. To achieve this, decoding the state-of-the-art is a prerequisite for further advancements.

While numerous systematic literature reviews (SLRs) have explored digital transformation in the built environment, most remain focused on specific technological domains or disciplinary boundaries. Recent contributions have primarily addressed sectoral integration processes [93,94] and the evolution of particular tools or technological paradigms [95]. Other studies focus on the application of digital advancements within narrowly defined contexts, such as heritage preservation [96], energy efficiency [97], asset management [98], distinct construction lifecycle phases [99], or circular economy frameworks [100]. In parallel, a growing number of bibliometric reviews have applied co-occurrence analysis to map research landscapes across these fields [101,102,103]. While these studies contribute to understanding thematic clusters within subdomains, their scope remains predominantly descriptive—visualizing keyword frequencies or co-authorship networks without cross-domain synthesis.

Unlike previous SLRs that provide overviews limited to individual technologies or disciplinary domains, this study introduces a cross-domain and cluster-based synthesis that integrates insights from AEC, UPD, and CH domains. While existing research has offered valuable but fragmented perspectives, an integrated methodological understanding connecting these domains across spatial and operational scales has remained largely absent. To address this gap, the present study employs a cluster-based SLR framework that combines bibliometric co-occurrence mapping with qualitative interpretation. This mixed approach enables the identification of methodological clusters and their interdependencies, revealing how diverse digital solutions co-evolve within a systemic, multi-level workflow for built environment observation. By decoding these clustered structures and their workflow relations, the paper advances the current state of research from descriptive mapping toward interpretative, cross-domain, and interoperable understanding of digital transformation—thus providing added value beyond existing literature.

Accordingly, the research questions are structured around three complementary dimensions:

• RQ1: What are the methodological clusters emerging from the intersection of digital advancements and the built environment, and how can they be visualized and interpreted?
• RQ2: What are the key features and attributes of identified solutions, and how do they group into functional categories?
• RQ3: What are the comparative roles and interdependencies of these clusters, and how do they inform potential for future research?

To ensure conceptual clarity, several key terms used throughout this paper are defined as follows: (1) the term “clusters” refers to the analytical categories identified through the co-occurrence and content analysis of solution-oriented studies, (2) the term “framework” represents the overarching methodological and conceptual structure derived from these clusters, used to interpret their systemic interrelations, (3) the term “platform” denotes the digital or operational environment where these methodological principles are applied, and (4) the term “solutions” describe specific digital applications, engines, or protocols documented in the literature that embody the practical outcomes of the identified framework.

The paper is structured as follows: The first part of the paper (Section 2) presents the materials and methods applied in this research, starting from elaboration on general research conceptualization based on the systematic literature review (SLR) and implemented in 4 synchronized research phases which are further explained separately: Data Search (Section 2.1), Data Selection (Section 2.2), Data Clustering and Co-occurrence (Section 2.3), and Comparative and Interdependency Analysis (Section 2.4). The second part of the paper (Section 3) presents the results of the clustering analysis. First, an overview introduces the four identified clusters, accompanied by the visualization of their distribution. Each cluster is then elaborated in detail: Cluster 1 (Section 3.1), Cluster 2 (Section 3.2), Cluster 3 (Section 3.3), and Cluster 4 (Section 3.4). The third part of the paper (Section 4) provides the discussion, organized in four dimensions. The first discusses the comparative overview of the four clusters (Section 4.1), while the second focuses on their interdependencies and systemic interplay (Section 4.2). Finally, Section 5 concludes with a synthesis of findings, outlining recommendations, and future directions for further research and advancement.

2. Materials and Methods

The research is based on the systematic literature review (SLR) supported by the visualization of the bibliometric networks through 4 synchronized phases that include (Figure 1): (1) data search—definition of the initial sample of literature documents, (2) data selection—preliminary screening and application of final criteria, (3) data clustering and co-occurrence—constructing and visualizing networks based on identified keywords, and (4) comparative and interdependency analysis—interpreting and comparing clusters across selected dimensions.

Thus, the SLR is performed (1) in a mixed review approach by both quantitative (bibliometric) analysis and qualitative (in-depth content) analysis, and (2) through utilizing Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) protocol, which is especially apt for mixed-method reviews that cover both quantitative and qualitative studies. The completed PRISMA 2020 Checklist is provided in the Supplementary Materials (27-item checklist). This hybrid SLR design was selected to move beyond purely quantitative bibliometric mapping, enabling the identification of conceptual and thematic clusters that expose underlying methodological patterns across the AEC, UPD, and CH domains.

2.1. Data Search

Data search was performed within the Scopus academic database through employment of the keywords research strategy (document search based on defined terms within the scope of the article’s title, abstract, and keywords—including Authors and Indexed Keywords). Scopus was selected as the primary source due to its extensive interdisciplinary coverage and consistent indexing of peer-reviewed journals and conference proceedings across the AEC, UPD, and CH domains. Having in mind the overarching objective of the study to approach the interrelation and knowledge integration of two notions—built environment and digital advancements, keywords were set in cross-cutting point of view: (1) from the perspective of built environment—(“built AND environment”) OR (“urban AND morphology”) OR (“urban AND typology”) OR (“urban AND form”) AND, (2) from the perspective of digital advancements—(“digital AND tools”) OR (“digital AND methods”) OR (“data AND visualization”). Several combinations of keywords were tested using Boolean operators (AND, OR, NOT) to combine keywords and refine the search, and after each iteration, an insight into the search results and their relevance was made (especially following the representation of literature documents by subject areas) (

Table 1. Keyword definition and refinement process.

Conceptual Dimension	Tested Terms	Decision	Rationale for Inclusion/Exclusion
Built Environment	urban space, city structure, urban fabric, urban area, spatial configuration	Excluded	Returned heterogeneous results dominated by geography, sociology, or urban policy papers without a methodological focus
Urban Form/Morphology	built environment, urban morphology, urban form, urban typology	Included	Provided stable linkage to research addressing spatial structure, typology, and form-related methodologies across AEC, UPD, and CH domains
	city form, urban pattern	Excluded	Although conceptually related to urban form, these terms produced highly variable results dominated by morphological or spatial-structure studies outside the digital-methodological scope
Digital Technologies	digital technologies, digital applications, digital processes, ICT tools, computational techniques	Excluded	Too broad or technical, produced outputs from computer science and information system domains, lacking relevance to spatial analysis and design
Digital Tools and Methods	digital tools, digital methods, data visualization	Included	Accurately reflected methodological and technical approaches central to digital observation and representation of the built environment
	spatial analytics, geospatial methods, 3D modeling, simulation modeling	Excluded	Produced similar document sets as the final included terms (digital tools, digital methods, data visualization), indicating redundancy without added conceptual specificity
Conceptual Synonyms	smart city, urban innovation, smart governance, digital transformation	Excluded	Although relevant thematically, these terms generated conceptual or policy-oriented literature outside the methodological scope of this review
Final Combined Search Strings	(“built AND environment”) OR (“urban AND morphology”) OR (“urban AND typology”) OR (“urban AND form”) AND (“digital AND tools”) OR (“digital AND methods”) OR (“data AND visualization”)	Final inclusion set	Balanced inclusiveness and precision, ensured coverage of multi-scale, cross-domain methodological studies suitable for cluster-based synthesis

). While Scopus ensures comprehensive representation of multidisciplinary research relevant to the study’s scope, it is acknowledged that future research could incorporate complementary databases to further enhance cross-validation of the results. During preliminary testing, additional searches were conducted on the Web of Science and IEEE Xplore to assess database complementarity. However, these provided overlapping or less methodologically structured results compared to Scopus, which was therefore retained as the primary database due to its interdisciplinary coverage across AEC, UPD, and CH domains.

The definition and refinement of the keyword set were conducted through a structured, multi-stage iterative process involving the multidisciplinary research team (11 members), whose expertise spans four core domains: (1) architecture (3 researchers), (2) urban planning and development (3 researchers), (3) urban sociology and behavioral studies (2 researchers), and (4) information technology and digital media (3 researchers). All team members are affiliated with university-based research environments, with professional experience ranging from early-career to senior researchers. The refinement procedure was implemented through six iterative rounds (

Table 2. Overview of the six iterative refinement sessions.

Iteration	Participants	Method	Main Actions	Outcome/Adjustment	Decision Criteria Used
Round 1	Full team (11 members)	Structured brainstorming and workshop	Initial generation of long-list terms across conceptual dimensions	Identified candidate terms and conceptual dimensions	Semantic breadth, disciplinary representativeness
Round 2	Subgroup/Domain representatives (4 members)	Test searches in Scopus	Conceptual dimensions testing	Identified mismatches between tested terms and methodological scope, removed terms yielding heterogeneous or non-methodological results	Methodological focus, domain relevance
Round 3		Comparative test searches	Conceptual dimensions testing	Eliminated terms producing unstable or overly narrow result sets, and refined the shortlist for combined testing	Precision, thematic alignment
Round 4		Cross-evaluation of results	Combined both term sets and tested Boolean operators	Identified optimal stability in terms involving form/morphology with tools/methods	Corpus stability, balanced subject-area distribution
Round 5		Focused Scopus evaluation	Narrowed to the most reproducible combinations	Final shortlist reached	Cross-domain coverage, multi-scale applicability
Round 6	Full team (11 members)	Consensus moderation	Final testing of the combined Boolean string	Final inclusion set confirmed	Reproducibility, conceptual clarity

). Two rounds were held as full-team workshops, while four rounds were conducted as focused expert consultations within the sub-group consisting of domain representatives. These sessions functioned as structured brainstorming and analytical testing cycles: in each iteration, candidate terms were jointly proposed, tested directly in Scopus using the same filter set, and evaluated according to pre-defined decision criteria.

Across the six iterations, alternative terms were systematically tested and evaluated according to the following criteria: (1) semantic precision, (2) methodological relevance, (3) domain coverage across AEC, UPD, and CH fields, (4) stability of returned document sets across temporal and subject-area filters, and (5) capacity to represent multi-scale spatial methodological work. Where tested terms generated heterogeneous, policy-only, or overly technical corpora, as documented in Table 1, they were discarded and replaced by more precise alternatives. From the perspective of the built environment, expressions such as “urban space”, “urban structure”, and “urban fabric” were considered but excluded because they generated heterogeneous results dominated by social and geographical studies without a direct methodological focus. Similarly, from the perspective of digital advancements, broader terms such as “digital technologies”, “digital applications”, and “digital processes” were tested but rejected due to their returning literature outside the scope of methodological frameworks (mainly policy studies). The selected combination proved optimal for capturing research explicitly addressing methodological innovation within AEC, UPD, and CH domains. In this iterative process, it was recognized that the most reference sample from a multiscale perspective is formed when using terms related to morphological characteristics and typologies of space in combination with general terms from the digital domain that enable the coverage of a wide spectrum of digital methodologies. Accordingly, the final keyword string represents the most stable combination, derived through cross-domain consensus and validated through the reproducibility of search results.

While the process did not rely on a single external standard, the adopted terminology is aligned with the distinction used in the DS4SSCC [42], which similarly differentiates digital tools as technical enablers and digital methods as structured procedural approaches guiding their application. In the context of built environment research, digital tools refer to technological instruments, software systems, and computational environments that enable the acquisition, processing, modeling, analysis, or visualization of spatial data [104], while digital methods, on the other hand, denote structured analytical or procedural approaches that employ these tools to produce, interpret, or communicate spatial knowledge [9]. While the two notions are conceptually distinct, they are interdependent in practice: (1) tools represent the technical enablers of methodological processes, whereas (2) methods define the frameworks through which digital tools are applied. Within this research, both terms are used in an integrative sense, encompassing related expressions found in the literature (such as digital applications, frameworks, engines, and platforms) to ensure broad interpretability and transferability of results across disciplines.

After the keyword research strategy, 5261 literature documents were obtained. Several literature selection criteria were further applied and technically justified as follows: (1) the language limitation to English documents was adopted to maintain consistency of terminology and metadata structure within Scopus and to facilitate accurate co-occurrence analysis, (2) the subject area selection covering “Social Sciences”, “Computer Science”, “Engineering”, “Environmental Science”, “Earth and Planetary Sciences”, “Arts and Humanities”, “Agricultural and Biological Sciences”, “Decision Sciences”, “Multidisciplinary” was defined to ensure inclusion of interdisciplinary studies addressing digital transformation in the built environment while excluding unrelated technical fields, and (3) the temporal scope from 2013 to 2025 was chosen to reflect the period of consolidated digitalization in AEC, UPD, and CH domains research following the emergence of smart-city paradigms across Europe and beyond. The initial search was conducted on 16 April 2025, while the keyword string was continuously verified until the finalization of the manuscript to include any newly indexed studies meeting the established criteria. With the aforementioned selection criteria, the selection was narrowed down to 2124 literature documents and exported to Scopus bibliographic database file (CSV format).

2.2. Data Selection

In the second research phase, a preliminary analysis and screening of content was conducted for 2124 literature documents through a top-down approach, starting with the title of the paper and keywords, then the abstract screening, followed by full-paper screening in cases where the title and abstract did not provide sufficient information to assess compliance with the inclusion criteria. Bearing in mind that the study is focused on the application of digital tools and methods for comprehensive observation of the built environment, the basic criteria for selection is threefold: (1) relevance to topic—literature that directly addresses the application of digital tools and methods for observing the built environment, covering relevant concepts, methodologies, technologies, and case studies, (2) research setting—applicability of the study in built environment related disciplines, i.e., the literature positioned in one of the professional domains of AEC, UPD or CH domains (including interdisciplinary and multidisciplinary perspective), and (3) methodological rigor—the literature that contains a clearly and systematically explained methodological procedure including study design, data collection techniques, and analysis procedures, to ensure the reliability and validity of the findings. Based on the conducted screening, the selection was narrowed down to 507 literature documents, and the initial Scopus bibliographic database file was reduced. The progressive refinement of the corpus is summarized in

Table 3. Screening and inclusion results.

Selection Step	Criterion Description	Sub-Criterion/Category	Number of Documents Meeting Criterion	Cumulative Total
Initial retrieval	Documents obtained after keyword search and Scopus filters (language, subject area, date range)	/	2124	2124
Step 1— Topic relevance	Direct focus on the application of digital tools and methods for observing the built environment	digital observation/representation	620	1462
		methodological development/frameworks	480
		integrative or comparative studies	362
Step 2—Research setting	Studies positioned in AEC, UPD or CH domains (including interdisciplinary and multidisciplinary scope)	AEC	460	911
		UPD	290
		CH	85
		Interdisciplinary and Multidisciplinary studies	76
Step 3—Methodological rigor	Papers with clearly explained methodological design, data-collection and analysis procedures ensuring reliability and validity of findings	/	507	507

, which shows the number of records meeting each inclusion criterion. This transparent breakdown illustrates how the initial dataset was systematically reduced from 2124 to 507 documents based on the defined selection logic.

The subsequent level of filtering aimed to distinguish between general contributions and those providing concrete digital solutions was performed within a twofold process. First, the 507 documents were examined with an emphasis on identifying explicit outcomes in the form of applications, platforms, engines, or protocols that operate within the built environment as an observatory framework. In this step, conceptual or general theoretical contributions were set aside in favor of solution-oriented studies. The application of the solution-oriented criterion focused on identifying papers that demonstrated the operational implementation or testing of digital tools and methods, presented verifiable outcomes through case studies, prototypes, or simulations, and provided methodological transparency allowing replication or transferability across AEC, UPD, and CH domains. Studies that discussed the digital transformation conceptually or without a concrete methodological articulation were excluded. Second, within this subset, a qualitative assessment was performed to evaluate the applicability and robustness of the presented solutions, focusing on whether the study demonstrates functional features, attributes, and operational roles of digital methods. To ensure the consistency and transparency of this evaluative process, a set of key performance indicators (KPIs) was applied, summarizing the principal criteria used to identify solution-oriented studies (

Table 4. Key performance indicators (KPIs) for identifying solution-oriented studies.

Indicator	Description	Example of Evidence in the Literature
Functionality	Demonstrated operation or implementation of a digital application, platform, engine, or protocol applied within a real or simulated context	Development of a web-based 3D city model, data processing platform, or visualization tool validated through case studies
Scalability	Potential of the solution to be adapted or generalized beyond the original study area, context, or scale	Frameworks or systems replicable across different urban or architectural environments
Interoperability	Integration and connectivity of the proposed digital solution with other systems or datasets	Solutions linking spatial databases or combining BIM and GIS data for multi-scalar analysis
Sustainability and Usability	Long-term applicability, accessibility, and user-oriented design supporting broader digital transformation goals	Approaches ensuring open data formats, user interaction, or sustainable performance monitoring

). Through this two-level selection, the dataset was narrowed to 29 literature documents, which constitute the final sample for analysis, as they not only meet the general inclusion criteria but also represent state-of-the-art, solution-driven research with direct practical implications. The complete list of these 29 documents, along with their evaluated characteristics, is provided in Appendix A to ensure transparency and reproducibility of the final selection.

2.3. Data Clustering and Co-Occurrence

To ensure methodological transparency and reproducibility, the data extraction and coding process was applied. The identification of features was conducted through inductive qualitative analysis of the final 29 solution-oriented papers, where recurrent terms, concepts, and operational characteristics were extracted as analytical features. In total, 30 distinct features were identified, each representing a recurring digital functionality, methodological attribute, or operational quality within the analyzed studies. Rather than being pre-defined or coded in binary form, features were frequency-based and content-driven, meaning that their inclusion in the co-occurrence matrix depended on their contextual relevance and number of occurrences across the literature corpus. Feature extraction was carried out collaboratively by three independent researchers, with reconciliation achieved through iterative cross-validation and consensus to ensure interpretive consistency. To verify the reliability of this coding process, inter-coder agreement was assessed using Cohen’s κ coefficient (κ = 0.86), indicating a high level of consistency among coders. While this procedure ensured internal reliability of the coding framework, we acknowledge that no external or independent validation of the clustering results, through alternative classification techniques, was conducted. This is addressed as a methodological limitation in Section 4.4, as such external validation would further strengthen the generalizability of the identified cluster structure. The resulting Feature Codebook, provided in Appendix B, defines each feature, illustrates its application through an example from the reviewed literature, and reports its frequency of appearance across the 29 papers. This codebook simultaneously constitutes the basis of the co-occurrence matrix.

On this dataset, a co-occurrence analysis was conducted to systematically explore interrelations between identified features and attributes. The analysis was performed using VOSviewer (version 1.6.20), a widely recognized software tool designed for constructing and visualizing bibliometric networks. VOSviewer enables the mapping of terms, keywords, and concepts based on their frequency of co-appearance, thereby supporting the detection of underlying patterns and clusters within the data. In this study, the tool was employed to detect conceptual proximities and interdependencies among the extracted solution features, applying the following parameters: (1) counting method—full counting, (2) normalization—association strength, (3) minimum occurrence threshold—0 (since the analysis included a closed set of 30 specific features derived from the literature, the objective was to establish relational proximity among all of them), and (4) layout—LinLog/modularity optimization (attraction = 2, repulsion = −1). Moreover, the minimal cluster size was defined through several configurations (ranging from 3 to 6 clusters) compared to assess network stability, link strength, and thematic coherence. When the minimum cluster size was set to 3, the resulting network produced a fragmented structure with several small, thematically overlapping clusters that lacked interpretative stability. In contrast, increasing the threshold to 6 generated overly aggregated clusters, merging distinct thematic domains, and obscuring methodological differentiation. The configuration based on a minimum cluster size of 4 achieved an optimal balance between network density and interpretative clarity, ensuring that each cluster represented a distinct methodological domain without redundancy or fragmentation. This structure was further validated through cross-verification by three members of the research team, confirming its conceptual robustness. These clusters were subsequently visualized in a matrix and analyzed both quantitatively and qualitatively. The quantitative aspect relied on the examination of links, total link strength (TLS), occurrences (OCC), and co-occurrence frequencies, while the qualitative interpretation focused on the thematic significance of the identified clusters and their alignment with broader research trends.

2.4. Comparative and Interdependency Analysis

Instead of conceiving the co-occurrence analysis as a purely quantitative exercise, the final methodological phase has been assessed in line with the quali-quanti tradition in digital sociology [105,106,107,108,109,110]. In this perspective, against treating the four identified clusters as self-evident statistical aggregates, these are rather interpretive entry points that require systematic comparison of their operational and conceptual roles. Such an approach enacts what Marres and Gerlitz [111] describe as the re-insertion of computational traces into sociological inquiry, transforming patterns into problems and numerical values into interpretive objects [112].

The comparative analysis was oriented toward examining the clusters across five key dimensions: (1) primary role in the workflow—assessing the position and function of the cluster within the broader digital workflow of built environment research and practice, (2) type of data and operations—classifying the nature of input data, modes of processing, and output operations characterizing each cluster, (3) level of user involvement—identifying the degree of human participation, expertise, and interaction embedded in the functionalities represented by each cluster, (4) typology of contribution and specialized role—distinguishing the clusters based on their targeted contributions or application-oriented deployment, and (5) relevant scale of application—mapping the clusters based to its corresponding operational scale. By structuring the comparison around these five analytical axes, the study enabled understanding of the distinctive characteristics of each cluster and their alignment with the wider research landscape.

In addition to the comparative overview, an interdependency analysis was carried out to capture the systemic relationships between the clusters. This analysis highlighted overlaps, complementarities, and potential synergies, illustrating how the clusters jointly contribute to a more integrated and comprehensive framework for digital observation of the built environment. Emphasis was placed on identifying both reinforcing and bridging roles among clusters, thus revealing pathways for their combined application and future methodological advancement. Together, these two complementary analyses—comparative overview and interdependency analysis—formed the interpretative framework for discussion, linking the technical results of the clustering process to their broader implications.

3. Results

The subject of the conducted clustering analysis is the identification of key features of digital solutions for built environment observation presented within the scope of the final sample. Content analysis indicated 30 unique keywords (phrases) that refer to key features that characterize the analyzed solutions. The interpretive synthesis builds directly upon the quantitative structure revealed through the co-occurrence analysis. The visualization produced by VOSviewer, based on OCC and TLS thresholds, served as a foundation for identifying four distinct yet interrelated thematic clusters (

Table 5. Review of clusters with the list of keywords that correspond to complete counts extracted from the final 29 solution-oriented papers, not binary presence/absence values.

Cluster Label (Color)/Topic	Keywords
Data Integration and User-Centric Analysis (Red Cluster)	collaborative modeling, Data Integration, dynamic visualization, evaluation and validation, indicator calculation, integration of spatial videos and geonarratives, interactive analysis, multiscale-analysis, quantification of change, semantic analysis, simulation and prediction, spatiotemporal analysis, user accessibility and public engagement, user behavior and perception analysis
Advanced 3D Spatial Analysis and Processing (Green Cluster)	3D spatial analysis, data acquisition, data interoperability, data processing, procedural modeling, querying and data retrieval, tiling and visualization
Real-Time Interaction and Digital Twin Support (Blue Cluster)	location-based analytics, modular functionality, real-time interaction, support for digital twin development, user interaction
3D Visualization (Yellow Cluster)	3D visualization and navigation, integration of GIS and HBIM, ownership and zoning analysis, typological classification

): (1) Data Integration and User-Centric Analysis (represented by red color, includes 14 keywords), (2) Advanced 3D Spatial Analysis and Processing (represented by green color, includes 7 keywords), (3) Real-Time Interaction and Digital Twin Support (represented by blue color, includes 5 keywords), and (4) 3D Visualization (represented by yellow color, includes 4 keywords). Quantitative metrics, including node centrality and connection density, indicated the relative significance and interdependence of key concepts within the dataset. These quantitative insights were subsequently cross-referenced with the content of individual papers to develop a qualitative understanding of how digital transformation in the built environment is methodologically and conceptually framed. In this way, the bibliometric mapping informed the narrative synthesis that follows, ensuring continuity between data-driven evidence and interpretive reasoning within the mixed-method design of this study. To strengthen the connection between the quantitative bibliometric indicators and the interpretive synthesis, a complete overview of all 30 analytical features with their corresponding OCC and TLS values, and cluster assignment is provided in Appendix C.

Figure 2 highlights “Data Integration” (Red Cluster) and “3D visualization and navigation” (Yellow Cluster) as central nodes that influence the relationships among clusters by acting as primary connectors. “Data integration” links analytical features (such as spatiotemporal analysis, semantic analysis, and quantification of change), facilitating interactions between “user accessibility and public engagement” (Red Cluster) and “user interaction” (Blue Cluster). On the other hand, “3D visualization and navigation” serves as the key bridge between spatial data representation features (Yellow Cluster) and data processing workflows (Green Cluster). Moreover, “user Interaction” (Blue Cluster) plays a pivotal role in connecting real-time engagement, and digital twin development ensuring interactive data retrieval and collaborative modeling. Finally, “3D spatial analysis” (Green Cluster) acts as the core analytical component, ensuring structured spatial data acquisition, procedural modeling, and interoperability before visualization. The specific position is occupied by “spatiotemporal analysis” as a linking feature which reinforces inter-cluster dependencies by allowing for temporal and multi-scale integration.

The hierarchical structure of the identified clusters reflects a systematic workflow of digital solutions for built environment observation. Data Integration and User-Centric Analysis (Red Cluster) functions as the core cluster, which integrates diverse datasets, enabling spatiotemporal analysis, and ensuring accessibility for user engagement. Supporting this foundation, 3D Spatial Analysis and Processing (Green Cluster) and Real-Time Interaction and Digital Twin Support (Blue Cluster) serve as supporting clusters—features within the Blue cluster provide structured spatial data processing through acquisition, interoperability, and procedural modeling, while the Green cluster enhances responsiveness through real-time interaction, modular functionality, and support for digital twin development. Finally, 3D Visualization (Yellow Cluster) operates as a specialized cluster, transforming processed data into spatial representations, integrating GIS and HBIM and providing the framework for typological and zoning analysis. This hierarchy of clusters underscores a structured interdependence, where Data Integration provides the foundational framework, spatial and interaction-based processes refine and analyze data, and visualization serves as the primary driver for interpretation.

3.1. Cluster 1: Data Integration and User-Centric Analysis

The core terms of this cluster are “Data Integration” (OCC = 13, TLS = 36) and “dynamic visualization” (OCC = 8, TLS = 26) followed by features of “user accessibility and public engagement” (OCC = 6, TLS = 19) and “spatiotemporal analysis” (OCC = 7, TLS = 19). Features affiliated with the subject cluster could be classified into 3 groups based on their specific purposes: (1) Data Integration and representation (Data Integration, dynamic visualization, integration of spatial videos and geonarratives), (2) analytical tools (spatiotemporal analysis, multiscale analysis, quantification of change, semantic analysis, evaluation and validation, indicator calculation, interactive analysis, user behavior and perception analysis), and (3) user-oriented features (user accessibility and public engagement, collaborative modeling).

At the core of the cluster, the dominant role of features related to Data Integration and representation (represented in 15/29 analyzed solutions from the final sample for in-depth analysis) is recognized. Content analysis indicates that Data Integration is performed for three-fold purposes: (1) to enable a framework for accurate and detailed 3D models development through combining Light Detection and Ranging (LiDAR) datasets, aerial images, and large-scale numerical maps [113] or through integration of various 3D geospatial data formats, such as City Geography Markup Language (CityGML) files and 3DCityDB relational schemas [114], (2) to enable a framework for synchronous understanding of different spatial and temporal scales [115], or (3) to enable unified datasets for further analysis through combining cadastral data (such as building height, construction year, and use) with urban planning regulations [116] or through combining satellite imagery, demographic data, and urban expansion indices to provide a comprehensive understanding of urbanization trends [117]. By its nature, it is a multi-source and multi-modal Data Integration that combines structured and unstructured data sources, for example: (1) integration of point clouds, satellite imagery, and BIM models, into a unified platform [118], (2) integration of CAD-based phased maps, historical records, and geospatial information, into an interactive format [119], and (3) integration of social media data, mobile app data, and survey data to capture physical, social, and environmental characteristics of urban spaces [120,121].

In terms of representation features, multiple tools and techniques are identified within the scope of dynamic visualization across spatial and temporal scales including (1) Level of Detail (LOD) techniques to enable efficient and dynamic loading of 3D models at different scales [118], (2) interactive tools, such as time sliders, to observe urban growth over different periods and to visualize spatial and temporal changes in land use and land cover [117], (3) application of lightweight and customizable rendering based on Web Graphics Library (WebGL) (Mapbox GL JS and Deck.GL) for scalable transitions across spatial levels, from local government units to wider regional areas [122], and (4) implementation of interface with map views, video players, and keyword filters, allowing users to visualize spatial trajectories and associated multimedia content [123]. Similarly, integration of spatial videos and geonarratives [123] is recognized as a feature that provides a visual analytics framework and accordingly, enables contextualized exploration of geographic locations.

Looking at the relations of features addressed within the scope of the subject cluster, it is recognized that Data Integration and representation features provide an operational framework for the further performance of a series of analytical tools. Spatiotemporal analysis is recognized as the dominant representative of analytical tools performed in order to (1) explore the historical evolution of the sites through layered visualizations representing different time periods [119], (2) visualize the current spatial distribution and structure of buildings and land uses using interactive maps, charts, and a time slider [124], (3) detect changes in building sizes, shapes, density, and street patterns [125] or changes in temporal patterns in human concentrations and activities [121] over different temporal sequences, and (4) analyze the temporal evolution of building stock, including changes in height, use, and compliance over time [116]. Complementary to the temporal dimension of analysis, multi-scale analysis is recognized in terms of providing a framework for observation of the built environment at various spatial levels [126], including cityscape, neighborhood, and individual building units, allowing users to explore the characteristics and development scenarios at multiple levels. The content analysis further indicated that the software framework of the analyzed digital solutions for the observation of the built environment provides the both features for the quantification of urban forms, including quantification of change [127,128,129] and indicator calculation [117], as well as qualitative observation mostly based on semantic analysis [121,123], and user behavior and perception analysis [120]. Finally, the interactive analysis feature is recognized as a connecting node in-between Data Integration and general analytical features towards dynamic visualization, enabling users to compare or analyze urban expansion patterns across spatial units [117].

In terms of user-oriented features, content analysis identifies user accessibility and public engagement as essential for attracting diverse audiences such as researchers, decision-makers, and the general public within both analytical and visualization tools [117,119]. These features serve multiple purposes: (1) aiding decision-making by offering insights for spatial development concepts and filtering sites by attributes [124,126], (2) supporting design through evidence-based insights into space usage and user feedback [120], and (3) facilitating public participation by improving communication of urban planning proposals, allowing citizens to visualize impacts and provide feedback [130]. Moreover, specific features are recognized in terms of collaborative modeling, which facilitates collaboration among users, enabling them to share, analyze, and discuss urban modeling data and results, as well as their validation [115].

3.2. Cluster 2: Advanced 3D Spatial Analysis and Processing

At the core of the second cluster is the feature of “3D spatial analysis” (OCC = 11, TLS = 15), which acts as a pivotal node towards features and refers to data management and computational processing. Features affiliated with the green cluster could be grouped as follows: (1) advanced analytical methods related to 3D spatial analysis, (2) features related to collecting and structuring data (data acquisition, data interoperability), (3) features related to processing and computation of data (data processing, procedural modeling), and (4) features related to data utilization (querying and data retrieval, tiling and visualization).

The first group indicates a pool of advanced analytical methods or protocols related to 3D spatial analysis, including (1) integration of specific tools for performing advanced spatial analyses, such as 3D intersection, clipping, and difference calculations [131], or model slicing, spatial relationships, and detailed inspection of spatial elements [118], (2) development of 3D information-rich city models to integrate geometric, semantic, and attribute data [132], (3) development of framework for 3D modeling at different levels of detail (LoD1 to LoD2) [113], and (4) implementation of automated workflows for advanced 3D change detection analysis [128].

In the context of the functionality related to collecting and structuring data, two specific solutions are recognized by content analysis. First is OSMnx, a Python package designed to streamline the processes of data acquisition and enhance data interoperability in street network analysis with three-fold advancement [133]: (1) automated data acquisition through facilitation of automated and on-demand downloading of a variety of geographic data, such as administrative boundaries, building footprints, and street networks directly from OpenStreetMap (OSM) with simple commands, (2) provision of multiple file formats enabling users to save street networks in various formats (e.g., shape files, GraphML, SVG), thus enhancing the interoperability of street network data across different applications and platforms including integration with GIS software, and (3) performance of topological and metric corrections to ensure that the street networks reflect accurate connections between nodes. The specificity of the second solution, 3D Geospatial Web Service, referred to as WFS3D (Web Feature Service for 3D) is recognized regarding [134] (1) integration of operations defined by the Open Geospatial Consortium (OGC) (such as GetCapabilities, DescribeFeatureType, and GetFeature) in terms of further promoting structured data requests and retrievals, and (2) creation of standardized 3D data models, particularly via the GIANT model including the integration of different geospatial data types such as topology and geometry into a unified structure within the PostGIS spatial database.

The concept of procedural modeling is recognized as the basis for the features related to data processing and computation. Key features and functionalities of procedural modeling are addressed within identified solutions (1) through the application of shape grammars that allow for the automatic creation of 3D models based on predefined rules [135], and (2) through the utilization of a technique known as extrusion, which generates 3D City Models (3DCMs) from 2D building footprint data provided by users [131].

Features related to data utilization can be grouped into two functional domains. The first domain focuses on querying and data retrieval, emphasizing the creation of advanced systems for interacting with 3D urban models. This includes the development of geospatial web services for visualizing, querying, and navigating 3D environments [134], the integration of spatial and temporal information into comprehensive 3D spatio-temporal GIS models [125], and the implementation of spatial semantic query engines specifically tailored for CityGML data structures [114]. The second domain pertains to tiling and visualization, addressing the challenges of efficient web-based management of large-scale 3DCMs. Techniques such as spatial tiling are employed to enhance the performance and usability of 3D content delivered via online platforms, as demonstrated in recent approaches to managing 3DCMs [131].

3.3. Cluster 3: Real-Time Interaction and Digital Twin Support

The third thematic cluster is centered around the core term “user interaction” (OCC = 8, TLS = 26). The features associated with this cluster can be organized into three sub-groups: (1) real-time data and analytics—which focus on decoding dynamic processes and interpreting what is happening within the system, (2) system architecture and scalability—which address how the system is structured and expanded to support complex functionalities, and (3) user-centric interaction—which concerns how users engage with access and manipulate digital environments.

The first group of functionalities within the blue cluster encompasses real-time interaction and location-based analytics characterized by modular functionality. The location-based analytics feature is exemplified by the Gateway web platform [136], which provides a user-oriented interface designed to support robust spatial analysis for non-technical users. The platform integrates four specialized modules: (1) Snapshot, for interactive map-based data exploration, (2) Predict, which employs statistical and ML models to forecast spatial trends, (3) Grade, enabling multi-criteria evaluation and ranking of geographic areas through weighted overlays or algorithmic scoring, and (4) Focus, which supports spatial decision-making by identifying optimal locations based on patterns of supply and demand. Two solutions for real-time interaction were identified: (1) the Gateway platform [136], already discussed, offers a cloud-based 2D interface with dynamic filtering, instant visual updates, and easy spatial analysis for non-specialists, and (2) a 3D visualization platform for subway engineering [118] integrates GIS, BIM, WebGL, and LOD techniques to allow seamless navigation of large 3D scenes and real-time queries directly in a web browser.

The second group of functionalities, focused on system architecture and scalability, encompasses the feature of support for digital twin development, identified in two approaches: (1) the integration of HBIM with GIS to generate semantically rich 3D urban models for detailed asset representation, lifecycle management, and decision making within a scalable GIS framework [137], and (2) the integration of multi-source spatial and temporal data through lightweight web technologies, enabling real-time visualization, dynamic urban monitoring, and scalable scenario simulation within a browser-based environment [122].

Finally, the core functionality of the cluster is centered on user interaction, supported by a range of integrated features that enhance dynamic spatial exploration and decision-making. These features can be broadly grouped into four functional domains. Temporal data interaction functionalities allow users to control historical phases by selecting layers representing different chronological stages in the site’s evolution [119], adjust time scroll-bars to dynamically visualize urban evolution across years or ranges [125], and explore animated temporal patterns such as human mobility flows (weekday vs. holiday patterns) for comparative scenario analysis [122]. Layering and comparative visualization functionalities enable users to switch between historical and current contexts by selectively displaying or hiding thematic data layers [119,122] and to directly compare two planning scenarios side-by-side using slider tools that convey spatial and temporal variability [130]. Interactive spatial exploration and data filtering functionalities include the ability to select and highlight specific map features, triggering tailored information pop-ups or dynamic feature highlighting [114,119,130], visualize data distributions through choropleth mapping filtered by administrative units and timeframes [136], conduct attribute-based filtering based on building characteristics such as function or construction date [134], and assign custom weights to data layers to prioritize geographic areas according to user-defined criteria [136]. Finally, navigation and scenario simulation functionalities allow users to virtually walk or fly through 3D urban models representing different historical periods [125,130], multi-scale exploration of the urban environment at citywide, neighborhood, and individual building scales [126], and simulate “what-if” scenarios by replacing existing buildings with new designs to immediately observe impacts on urban form indicators such as landmark visibility and sky exposure [126]. Collectively, these features illustrate a comprehensive and user-centered approach to engaging with complex urban datasets across temporal, spatial, and scenario-based dimensions.

3.4. Cluster 4: 3D Visualization

The fourth thematic cluster is anchored by the core term “3D visualization and navigation” (OCC = 14, TLS = 30), which represents the substantial advantage over static visualizations. In addition to this central feature, the cluster comprises three interconnected concepts that establish internal relationships and links to other clusters. “Integration of GIS and HBIM” is positioned autonomously within the cluster, directly linked to the core term and highlighting its integrative and enrichment role by connecting geometric models with descriptive and semantic datasets. “Ownership and zoning analysis” functions as a bridging node towards the “dynamic visualization” cluster, while “typological classification” serves as a connecting feature to “spatiotemporal analysis,” together forming analytical layers embedded within the 3D context. Analyzing the functionality of the selected solutions, a number of individual features associated with the 3D visualization and navigation were recognized, including representation aspects, semantic-interactive exploration, navigation, and feature selection, as well as different modes of analysis.

The dominant feature of the majority of analyzed solutions is the realistic city environment representation [113,114,132,134,138], including specific advanced methods such as view frustum culling (only rendering what is visible) and tessellation shaders (maintaining interactive frame rates, even when visualizing complex cityscapes in 3D) [126]. A particularly significant advancement recognized in the domain of representation refers to the derivation of 3D building heights, where generating Digital Surface Models (DSMs) and Digital Terrain Models (DTMs) from the stereo satellite imagery directly calculates building heights [129]. This approach addresses one of the major challenges in 3D urban modeling—the accurate definition of the third spatial dimension and, consequently, the representation of building volume.

Within the scope of semantic-interactive 3D exploration, the recognized solutions mainly integrate advanced tools for linking spatial, semantic, and attribute data through an interactive 3D environment. The interactive 3D search functionality is recognized in two forms: (1) one combining a searchable legend with two-way navigation between the model and its metadata, enabling users to locate specific terms or features directly within the 3D model [130], and (2) another where selections or filters applied in the data exploration view—using parallel coordinates charts or tabular interfaces—are dynamically reflected in the 3D visualization [126]. Complementing this, attribute and semantic data access functionalities are recognized in solutions that either connect attribute-based urban datasets to the physical forms of buildings and their surroundings, allowing users to (1) select specific elements in the 3D view to retrieve associated attributes and semantic information [126,134] or (2) harmonize semantic compatibility between cadastral geometries (actual built forms) and planning regulation geometries (legal forms) [116]. Furthermore, ontology-driven highlighting is recognized as a feature that ensures that the results of semantic or attribute queries are spatially explicit, visually emphasizing relevant objects within the 3D environment [114].

In the reviewed literature, a set of functionalities related to navigation and feature selection has been recognized, emphasizing different modes of interaction within 3D environments: (1) enabling users to directly select and interact with 3D features such as buildings or roads [114], or to access pop-up attributes (e.g., maximum height, usage, area) [130] by clicking on elements in the scene, (2) supporting advanced navigation through intuitive camera controls that allow movement from street-level to aerial perspectives [126,130], pedestrian-centric visualization [116], above–below ground perspective switching [46], or precomputed viewpoints integrated into automatic camera paths [139], (3) applying multi-resolution approaches that enable navigation across different scales—from city-wide or neighborhood levels down to individual buildings [126], and (4) extending navigation beyond spatial movement to topological exploration, where each network element is represented both geometrically in 3D and within a network-graph structure [46].

Beyond general navigation and search functionalities, the literature also points to specific modes of visualization and interaction that support scenario evaluation and regulatory analysis: (1) comparison or slider mode, where tools allow users to view and contrast two different scenarios [130], or to simulate floor space index scenarios that dynamically illustrate changes in building volume, density, and overall massing through 3D navigation [132], and (2) overlay mode where systems superimpose 3D representations of existing buildings with planning regulation models, using color-coded indicators to show conformity status (e.g., red for overbuilt, green for underbuilt, gray for compliant) [116].

4. Discussion

4.1. Comparative Overview of Four Identified Clusters

The previously presented content analysis, structured in accordance with the visual network of clusters (nodes and links that constitute this network), enabled the identification of both overarching functionalities and specialized features or attributes associated with the analyzed solutions. The review of solutions indicated that the identified clusters play specialized roles (typology of contribution), while their corresponding features and attributes (1) take on distinct positions within the workflow, (2) perform different operations and engage with specific types of data, (3) operate across particular spatial scales within the multi-scale framework, and consequently (4) involve varying levels of user engagement and interaction (

Table 6. Comparative overview of four identified clusters, presented in sequential order that corresponds to the workflow stages. This structure reflects the functional logic visualized in Figure 2.

Cluster	Primary Role in the Workflow	Type of Data/Operations	Level of User Involvement	Typology of Contribution (Specialized Role)	Relevant Scale of Application (Multi-Scale Linkage)
Cluster 1: Data Integration and User-Centric Analysis	Input foundation—provides unified, multi-modal datasets and analytical context	Semantic and multi-scale (textual, numerical, multimedia sources)	High involvement of stakeholders and the general public (public engagement, participation)	Analytical–fundamental contribution—provides the basis for all subsequent phases	Operates across all scales (object–building–neighborhood–city) by integrating diverse data sources and engagement layers
Cluster 2: Advanced 3D Spatial Analysis and Processing	Processing stage—transforms raw and integrated data into structured, analyzable formats	Geometric and semantic (3D models, networks, procedural models)	Predominantly expert tools (analytical focus)	Technical–analytical contribution—enables robust data processing and computations	Predominantly at building and neighborhood scales where geometric precision and semantic enrichment are essential
Cluster 3: Real-Time Interaction and Digital Twin Support	Dynamic interaction—enables real-time engagement, simulation, monitoring, and decision support	Dynamic (real-time, location-based, scenario-driven)	Combination of expert use and user interaction (interactive tools, participation)	Operational–interactive contribution—enables simulations, monitoring, and decision support in real time	Building and district levels, supporting modular digital twin systems, simulations, and operational feedback loops
Cluster 4: 3D Visualization	Output and communication—transforms processed data into interpretable spatial representations for experts and the public	Visual (3D representations, scenarios, overlays, GIS–HBIM integrations)	High accessibility for a broad spectrum of users (from experts to the public)	Communicative–interpretative contribution—translates complex data into comprehensible spatial representations	Neighborhood and city scales, providing communicative and decision-support environments for policy-making and public participation

).

The comparative synthesis has revealed a sequential workflow articulated through the four clusters, ranging from input (Data Integration and User-Centric Analysis) to processing (Advanced 3D Spatial Analysis and Processing), interaction (Real-Time Interaction and Digital Twin Support), and finally output (3D Visualization). It has been recognized that Data Integration and digital-twin platforms form the initial input layer by harmonizing heterogeneous datasets and transformation pipelines (CSV/GeoJSON/LAS to JSON/GeoJSON/binary) [122]. Within this stage, map-based tactical dashboards have been identified as instrumental for forward-looking settlement scenarios [124], while user-friendly analytics modules have been acknowledged as critical for enabling engagement of non-experts [136]. These inputs have been shown to feed advanced spatial models that structure urban change across time [125], enrich semantics via GeoSPARQL over CityGML [114], and formalize network-level analysis [133]. The analysis further indicates that model outputs are subsequently recognized as enablers of interactive, real-time, and modular operations on large-scale 3D city models. This has been demonstrated through the use of tiling and procedural modeling aligned with OGC standards [131] and predictive planning scenarios [140]. Finally, the visualization cluster has been identified as the communicative endpoint of the workflow, translating analytical results into accessible media. This role has been observed in the form of domain-specific real-scene 3D visualization [118], heritage time-maps [119], and WebGIS interfaces enabling temporal land-change exploration [117].

The cross-cluster comparison has revealed an ascending gradient of data and operational complexity. It has been recognized that initial stages mobilize semantic and multimodal inputs (e.g., demographics, feature/flow/point-cloud data) with explicit transformation pipelines [122]. This baseline has been shown to evolve into geometric–semantic 3D models and query structures [114,125], before extending toward dynamic, real-time and location-based analytics [124,131], and ultimately communicative, scenario-oriented representations [117,118,119]. It has also been identified that operations span a wide methodological spectrum, ranging from network construction and metrics such as OSMnx [133], to procedural city modeling derived from cadastral or service datasets in 2D (CityEngine) [127,135]. The comparative review has further highlighted the role of volumetric and FSI-style scenario simulations [132] and HBIM–GIS integrations as pathways toward higher levels of detail and semantic richness in testing scenarios [137,139]. Finally, it has been emphasized that platform-level interoperability is systematically reinforced through standards (CityGML, 3D Tiles) and web-based technology stacks (PostGIS, Cesium). These frameworks have been recognized as essential in sustaining a vertical interoperability continuum, enabling the smooth transition from raw data inputs to communicative, user-facing outputs [117,131].

The comparative analysis has highlighted that user involvement is unevenly distributed across clusters and phases, shifting according to both the purpose and the operational layer. At the input stage, it has been recognized that stakeholder and public participation are foregrounded, with emphasis on accessibility through non-expert analytics workflows and exploratory dashboards [124,136]. The processing stage, by contrast, has been shown to remain primarily expert-oriented, requiring technical proficiency for spatio-temporal modeling, semantic querying, and network-level analysis [114,125,133]. It has further been observed that real-time and digital-twin environments introduce hybrid modes of engagement, in which expert and non-expert users interact through scenario control, modular function design, and large-model operations [122,131]. Visualization clusters have been recognized as broadening access once again, enabling heritage-based and spatio-temporal storytelling as well as web-based change exploration, thereby lowering barriers to participation [117,119]. Finally, it has been underscored that design-support platforms integrate behavioral and environmental evidence to inform adaptive interventions, effectively linking professional expertise with community-driven insight [120,123].

The comparative synthesis has confirmed that the identified clusters perform specialized yet complementary roles, forming a polyfunctional ecosystem. At the analytical–fundamental level (input), it has been recognized that integration and transformation layers, along with twin-ready data, establish the evidentiary base required for higher-order operations [122,124]. The technical–analytical stage (processing) has been observed to consolidate computational methods for structuring and enriching urban dynamics, from spatio-temporal 3D GIS and semantic querying over CityGML to procedural 3D generation and HBIM–GIS couplings, thereby formalizing the analytic backbone of the workflow [46,114,125,127,133,135,137,139]. It has further been highlighted that the operational–interactive layer (real-time) introduces functions critical for decision support, such as large-model handling, standard-based tiling pipelines, and interactive dashboards that enable simulation, monitoring, and prioritization [124,131,132,136,140]. Finally, the communicative–interpretative role (output) has been underscored as essential in translating technical complexity into actionable spatial narratives, ranging from 3D domain-specific visualization to heritage time-maps and web-based change viewers [117,118,119]. The ecosystem thus reflects not only technical stratification but also a methodological synergy, where each role reinforces the next in a vertically integrated workflow.

4.2. Interdependencies Between Identified Clusters

At the same time, the analysis highlighted that the full potential of digital solutions is not achieved in isolation, but rather through the interplay and complementarity of clusters. The interdependencies between them point to a systemic character, where synergies across functionalities provide opportunities for comprehensive integration of tools and methods.

Table 7. Interdependencies between identified clusters.

Cluster	Cluster 1: Data Integration and User-Centric Analysis	Cluster 2: Advanced 3D Spatial Analysis and Processing	Cluster 3: Real-Time Interaction and Digital Twin Support	Cluster 4: 3D Visualization
Cluster 1: Data Integration and User-Centric Analysis	—	Provides multi-source datasets (LiDAR, cadastral, social media, geonarratives) that enable 3D spatial analysis, interoperability, and procedural modeling	Supplies integrated spatiotemporal and semantic data necessary for real-time interaction, modularity, and digital twin development	Acts as a feeder of unified datasets (historical, cadastral, demographic) that visualization tools transform into communicative outputs
Cluster 2: Advanced 3D Spatial Analysis and Processing	Requires integrated and harmonized data sources (semantic, temporal, spatial) from Cluster 1	—	Produces structured outputs (procedural models, queries, data retrieval results) that support interactive simulations and dynamic monitoring	Generates analytical outputs (change detection, LoD models, indicators) that are translated into visual representations and overlays
Cluster 3: Real-Time Interaction and Digital Twin Support	Dependent on accessible and harmonized Data Integration for enabling live analytics and modular functionality	Builds on 3D analytical outputs (interoperability, queries, advanced processing) to power scenario-based simulations and twin models	—	Strengthens visualization through dynamic filters, slider comparisons, overlaying planning regulations, and interactive highlighting
Cluster 4: 3D Visualization	Needs semantic and temporal integration to ensure meaningful visual representation (e.g., zoning + cadastral data)	Relies on outputs of 3D analysis for realistic rendering (geometry, LoD, analytical slicing)	Requires interactive and dynamic features (real-time updates, scenario comparisons) to enhance communicative role	—

expands upon the conceptual framework established in Table 6 by detailing the interdependencies between clusters. The matrix illustrates both directional and reciprocal relationships, highlighting how data flows across different phases of the workflow from the input stage (Cluster 1) to processing (Cluster 2), interaction (Cluster 3), and output (Cluster 4), and how feedback loops reinforce the continuity of digital processes.

The analysis has recognized that Data Integration platforms represent the indispensable input layer on which all other clusters depend. Their role is not only in harmonizing heterogeneous sources (CSV, GeoJSON, LAS) into usable formats (JSON, binary) [122], but also in framing urban foresight through dashboards and simulation interfaces [124]. The review confirmed that these platforms rarely end in themselves; rather, their value lies in providing accessible, multi-format datasets that enable processing, twin development, and ultimately visualization. Even proprietary solutions like Gateway [136] show how integration doubles as a bridge to non-expert user groups, thus mediating between technical complexity and broader application. In the context of interoperability, this cluster directly aligns with open geospatial standards such as OGC Web Services and GeoSPARQL, which ensure semantic connectivity and consistent data querying across heterogeneous urban datasets.

Building on this foundation, the analysis highlighted that advanced 3D spatial analysis associated with the second cluster is critically dependent on harmonized datasets from the integration cluster. Yet it simultaneously produces structured outputs essential for downstream functions. Studies underline how spatio-temporal GIS models [125] and semantic enrichment engines [114] transform raw inputs into spatio-semantic knowledge, while network analytics [133] and procedural modeling [127,135] extend interpretative capacity. This review further noted that these outputs feed directly into both simulation and visualization clusters, producing change detection, LoD models, and other indicators that become communicative assets when rendered. In this way, 3D spatial analysis emerges as both technically demanding and structurally generative. This cluster aligns with CityGML and CityJSON standards for semantic 3D city modeling, as well as 3D Tiles and glTF specifications that support large-scale streaming, tiling, and visualization of complex urban geometries.

The discussion also revealed that analytical outputs from the second cluster serve as the backbone of real-time interaction and digital twin support within the third cluster. Real-Time Interaction and Digital Twin Support cluster relies on harmonized input and 3D processing results to operationalize modular scenarios and dynamic monitoring environments. The analysis recognized predictive dashboards [124] and large-scale tiling/procedural workflows [131] as illustrative of how static data gains new value through live updates, scenario-based simulations, and twin-ready models. Its operationalization strongly corresponds to interoperability standards, such as the ISO/IEC IoT frameworks, enabling continuous bidirectional communication between physical and digital systems.

Finally, the analysis emphasized that visualization platforms, though often the most accessible entry point for diverse users, are fully contingent on prior cluster contributions. Without semantic and analytical layers, visualization would remain superficial, yet with integrated features, it becomes an interpretative and communicative device. WebGIS tools [117] and scientific 3D web services [134] demonstrate how outputs from upstream clusters are transformed into interactive media, while advanced 3DCM management solutions [131] underscore the importance of standards for scalability. As such, visualization emerges not merely as a representational endpoint, but as a communicative layer that synthesizes the outputs of all preceding clusters. This aligns with 3D visualization standards such as glTF, Cesium 3D Tiles, and WebXR, which ensure web-based accessibility, interoperability, and real-time rendering of spatial content across diverse user environments.

4.3. Beyond the State of Research

The four identified clusters reflect a methodological structure that closely parallels the evolving global and European frameworks driving the digital transformation of the built environment. The first cluster, Data Integration and User-Centric Analysis, resonates with frameworks that institutionalize interoperable and transparent data governance across the entire lifecycle of the built environment. This dimension aligns with the Directive 2014/24/EU on Public Procurement (2014) [38], the Energy Performance of Buildings Directive (EPBD—Revised) (2024) [39], and the Construction Products Regulation (2024) [40], which collectively introduce BIM, Digital Building Logbooks, and Digital Product Passports (DPP) as standardized mechanisms for ensuring data traceability and efficiency. Comparable global initiatives such as the United for Smart Sustainable Cities (U4SSC) platform (2025) [34], the ITU-T Study Group 20 (2024) [141], and the UN-Habitat People-Centered Smart Cities Program (2025) [33] further emphasize the importance of interoperability, data ethics, and user-centered approaches in the digital transformation of urban systems.

Building upon these foundations, the second cluster, Advanced 3D Spatial Analysis and Processing, corresponds to the innovation-oriented trajectory of the European Green Deal (2019) [35], the EU Industrial Strategy and Construction Transition Pathway (2023) [37], and the Renovation Wave Strategy (2020) [142], which position digital technologies as enablers of climate neutrality, circular economy practices, and performance monitoring. Within this policy ecosystem, Horizon 2020 and Horizon Europe projects such as DigiPLACE (2020) [41] and DigiBUILD (2023) [143] advance frameworks for simulation-based design, predictive modeling, and Data Integration at multiple scales. These objectives parallel the World Economic Forum’s Fourth Industrial Revolution for the Built Environment (2022) [144], which highlights the fusion of physical and digital systems through AI, sensor technologies, and blockchain, promoting adaptive environments.

Furthermore, the third cluster, Real-Time Interaction and Digital Twin Support, connects directly to frameworks that emphasize continuous data flows and dynamic feedback. This direction is strongly embedded in the UN Sustainable Development Goal 11 (Sustainable Cities and Communities) (2015) [2], the SDG Digital Acceleration Agenda (2023) [32], and the ITU Digital Transformation Master Plan for Smart Sustainable Cities (2023) [145], which together promote real-time monitoring and scenario-based urban management. Within the European policy landscape, these principles are operationalized through the EU Local Digital Twins Toolbox (2025) [43] and the European Data Space for Smart and Sustainable Cities and Communities (DS4SSCC) (2022) [43], which enable cross-sectoral data sharing and model-based governance at the urban and regional scales. The EU Climate Adaptation Strategy (2021) [146] and the Smart Cities Marketplace (2025) [147] further reinforce the role of digital twins in supporting climate resilience and smart urban operations.

Finally, the cluster 3D Visualization corresponds to the communicative and participatory dimension of digital transformation, where visualization acts as a bridge between data complexity and public understanding. Frameworks such as the New Leipzig Charter (2020) [36], the New European Bauhaus Initiative (2021) [148], and the Urban Agenda for the EU (Pact of Amsterdam) (2016) [149] highlight the importance of inclusive, human-centered, and design-driven digitalization, while the UN New Urban Agenda (2017) [31] promotes the use of digital tools to support accessibility, transparency, and community engagement. This vision is reinforced by the Living-in.EU Movement (2020) [150], which advocates for shared digital infrastructures and open interoperability standards to enhance citizen participation and co-creation.

Taken together, these policies correspond to the position of the four clusters as complementary components of a unified digital ecosystem spanning from interoperable Data Integration and high-performance 3D analytics to dynamic twin infrastructures and visualization. In addition to revealing methodological interdependencies, the four-cluster framework provides a practical foundation for developing digital transition pathways in the built environment. Its structure can inform municipal digital strategies by guiding how cities design interoperable data ecosystems that connect planning, infrastructure, and citizen engagement tools. In the field of heritage management, the framework offers a roadmap for integrating documentation, monitoring, and visualization systems through shared spatial databases and digital twins. For smart city governance, the findings suggest how multi-scale Data Integration and real-time interaction solutions can enable participatory, and evidence-based decision-making. By linking these methodological domains with operational and governance mechanisms, the study contributes to advancing the practical implementation of digital transformation policies and strategies.

4.4. Limitations of the Study

While the present study provides a systematic synthesis of digital solutions for multi-scale built environment observation, several limitations should be acknowledged. First, the review was conducted using a single-database pipeline (Scopus), which, although comprehensive and interdisciplinary, may exclude relevant studies indexed elsewhere. Second, the feature extraction process was based on manual, inductive coding, which—despite high inter-coder agreement (κ = 0.86)—introduces a degree of interpretative subjectivity. Third, although the clustering structure was algorithmically generated and internally validated through high inter-coder agreement, an external validation of the cluster configuration was not undertaken. Future work should incorporate independent expert review or alternative clustering techniques to reinforce the stability and transferability of the identified methodological groupings. Fourth, the absence of a formal risk-of-bias or quality appraisal protocol tailored specifically for solution-oriented studies may have limited the systematic evaluation of methodological robustness across the included works. Finally, the distribution of sources indicates a potential venue bias toward conference proceedings, particularly within technology-driven clusters, which could affect the balance between conceptual and applied contributions. These limitations do not undermine the validity of the findings but rather delineate the scope of the current synthesis and highlight directions for future research, including database expansion, automation of feature extraction, and development of dedicated appraisal frameworks for solution-based studies in the built environment domain.

5. Concluding Remarks

This study set out three guiding research questions that framed the exploration of digital tools and methods in the context of built environment observation.

(RQ1) The first question addressed the extent and scope of available literature, testing whether bibliographic evidence could confirm the relevance of the field. The conducted SRL revealed a broad framework of bibliographic units (2124 documents), thereby confirming the thematic relevance and ongoing academic interest. However, only a very small proportion (0.55%) of literature documents from the initial data search demonstrated direct applicability to practical solutions. This indicates that the field remains largely conceptual, with theoretical foundations under continuous examination, while practical contributions are still at the stage of prototype presentation or technological validation. The progressive narrowing of the literature corpus, as illustrated in the data selection, shows that each refinement step applied more stringent criteria related to topic relevance, disciplinary positioning, and methodological rigor—ultimately isolating only 507 documents with replicable methodological design, and from these, 29 that demonstrated a solution-oriented framework. Therefore, this low percentage does not reflect the limitation of the scope, but the methodological selectivity of the review, which intentionally emphasized high thresholds of transparency, reproducibility, and demonstrated functionality. When assessed through key performance indicators, most studies failed to reach the ‘solution-oriented’ level, underscoring the current imbalance between conceptual research and practically validated digital approaches in the built environment domain.

(RQ2) The second question focused on the nature of methodological clusters that characterize the identified solutions. Through co-occurrence and clustering analysis of the final sample (29 documents), four distinct clusters were identified, each representing specific configurations of features and attributes. These clusters revealed orientations toward different stages of workflow (from data acquisition and modeling to visualization and integration), highlighting the fragmented nature of digital solutions. Although robust in their individual focus, these solutions remain specialized rather than holistic, each addressing particular tasks or phases without offering integral coverage of the entire workflow. Furthermore, this gap may also reflect the limited integration between academic research and the innovation sector, where the scientific framework is insufficiently informed by technological advancements and industry-driven development.

(RQ3) The third question examined systemic interplay and interdependencies between clusters. Comparative analysis across four categories—primary role in the workflow, type of data and operations, level of user involvement, and typology of contribution -demonstrated overlapping potentials and possible complementarities between clusters. The interdependent analysis suggested that, while individual solutions may appear partial, their combined perspectives can point toward a more integrated digital framework for observing the built environment.

In this respect, the review highlights an important research gap but also a future opportunity—current solutions provide valuable building blocks, yet their integral consolidation into a comprehensive digital platform remains an open challenge. The findings suggest that advancing the field requires moving beyond isolated functionalities and toward designing architectures that synthesize clusters into systemic, interoperable frameworks. Such an integral approach would enable the digital observation of the built environment to evolve from fragmented prototypes into fully operational platforms relevant for supporting decision-making, planning, and knowledge generation. Looking ahead, several practical and research-oriented directions emerge from this study: (1) future efforts should focus on developing interoperability standards that enable consistent data exchange across BIM, GIS, IoT, and digital twin ecosystems, ensuring that digital workflows are technically compatible and scalable, (2) pilot studies that combine multiple clusters identified in this research could demonstrate their complementarity and reveal operational trajectories for integrated implementation in UDP, AEC, and CH domains, and finally, (3) establishing open-source digital observatories could facilitate transparent data sharing, long-term monitoring, and collaborative innovation across municipalities, academic institutions, and private sectors.