Analyzing Visitor Behavior to Enhance Personalized Experiences in Smart Museums: A Systematic Literature Review

Abstract

This systematic review provides an analysis of information gathered from 33 chosen publications during the past decade. The analysis reveals the primary methodologies applied and identifies the visitor behaviors that enable personalized content delivery. Statistical and Data Analysis is the predominant methodology in the reviewed publications. The methodology is present in 97% of the publications. AI and Machine Learning (63.6%) and Mobile/Interactive Technologies (60.6%) are most frequently paired with this methodology. Behavioral Analytics Platforms and Mobile/Wearable Devices are the most used technologies (42.4%) for delivering personalized content. A total of 39.4% of publications utilize Location Tracking Systems. The most frequent visitor behavior analysis focuses on Interactive Engagement and Movement Patterns, which occur 72.7% of the time, before Learning Patterns and Physical Positioning, which occur 63.6% of the time. The behavioral analysis of Group Dynamics (27.3%) and Emotional Response (18.2%) represents the least common practice when museums personalize their content despite the significance of social interaction analysis among visitors. The leading content personalization methods currently include real-time personalization systems combined with AI-driven systems and location-based technologies. Personalized content delivery systems face challenges including privacy protection and scalability issues paired with expensive implementation costs, which especially affect smaller museums. Researchers should explore how new technologies, such as virtual reality, augmented reality, and advanced biometric systems, can be integrated into future developments.

1. Introduction

1.1. Background and Context

Museums have been essential throughout history for safeguarding cultural heritage while simultaneously providing educational experiences and promoting community involvement. Museums function as active venues that unite historical artifacts with artistic expression and shared cultural memories. The development of the “smart museum” concept started as a solution to traditional museum experiences, which consisted of static displays and few interactive elements because visitors received information through one-way communication by absorbing pre-selected content. The digital revolution has thoroughly changed this model by creating new possibilities for how museums interact with visitors [1]. Smart museums developed through three separate stages as they advanced. Between 1990 and 2005, museums began digitizing their collections while starting to create simple multimedia exhibits. During the second phase, between 2006 and 2015, museums adopted interactive installations alongside mobile applications to initiate digital engagement with visitors. From 2016 to now, smart museums have significantly enhanced their capabilities through advanced artificial intelligence (AI) systems combined with real-time analytics and platforms that deliver personalized visitor experiences [2].

The complexity of content personalization in smart museums arises from needing to integrate theoretical models from multiple disciplines. Visitor experiences can be effectively analyzed and enhanced through the application of cognitive psychology models along with learning science research and human–computer interaction theories. This section covers major models and their usage within smart museums.

Falk and Dierking’s contextual learning model [3] describes museum visits as outcomes of interactions between three primary dimensions, which include personal aspects, social influences, and physical surroundings. The personal dimension examines the visitor’s interests, knowledge base, and motivations, which together influence their previous experiences. Smart museums utilize personalization techniques that analyze visitor preferences and behavior patterns to create content that responds directly to these individual factors. The social dimension looks at how group dynamics and social interactions form crucial parts of the museum visit. Through interactive technologies and installations, museums provide platforms for collaborative learning and discussions which embody social learning theory. The physical dimension emphasizes how environmental factors shape the interaction space. Visitor experiences become more dynamic through location tracking technology combined with geo-targeted content, which enhances spatial visit elements while producing personalized and flexible routes.

The self-determination theory from Deci and Ryan [4] provides important motivational concepts which have significant relevance to personalization. Smart museums can enhance visitor engagement through the implementation of autonomy, competence, and connectedness principles. Personalized content systems, which enable users to select their own itinerary or themes, strengthen their sense of autonomy. The evaluation of user engagement with interactive displays and assessment feedback for learning targets strengthens competence while supplying personalized advice.

According to Piaget’s constructivist theory [5], students learn best when they actively engage in the educational process. Visitors engage with dynamic learning content through interactive technologies like virtual reality and intelligent displays so that they become active participants rather than passive observers. These technologies advance cognitive development and content comprehension through context-based challenges and tasks.

The technology continuum defined by Davis et al. [6] provides the foundation needed to assess how museums can integrate artificial intelligence (AI) and virtual reality (VR) technologies. Smart museums need to evaluate utility and usability to guarantee successful technology adoption among diverse audiences. AI generates personalized suggestions and analyzes visitor behavior to enhance traditional methods, thereby creating an effective and adaptive connection between visitors and content.

Smart museums provide several advantages through personalized content yet require careful attention to potential challenges. Personalized recommendations serve as cognitive development “scaffolds”, yet technological dependency risks diminishing social interaction importance, as identified by Falk and Dierking [7]. Preserving the cultural and social context while offering personalized content poses a significant challenge.

Smart museums that incorporate theoretical models not only improve visitor experiences but also deepen their analysis of how technology interacts with cultural content and human behavior.

Personalization defines the smart museum concept by reflecting diverse visitor backgrounds and interests along with their learning methods. Personalization technologies enable museums to create experiences that respond to the needs of each visitor. Museums create personalized experiences through content adaptation according to visitor expertise levels while providing interest-based route recommendations and interactive element adjustments for different learning styles. Contemporary museums evolve into interactive environments where advanced technology delivers personalized and impactful visitor encounters [8]. The transformation goes beyond simple technological use because it represents a core shift in museum spaces operating as complex systems that enhance human connections through cultural learning [9,10].

Museum institutions gain advantages from technological innovations that break through physical and cognitive limits to offer visitors improved ways to interact with and comprehend cultural collections. Artificial intelligence drives museum transformation by enabling the production and management of data to tailor visitor experiences [11]. Researchers obtain precision in understanding visitor behavior and cognitive processes through computer-based technologies like AI and machine learning combined with sensor networks, which enable real-time adaptation to exhibition experiences. Advanced AI-based projection mapping and augmented reality (AR) tools create complex experiences that engage multiple audience demographics, according to Ribeiro et al. [12]. Contemporary technologies cause a basic transformation by shifting museum visitors from passive observers to active participants.

Smart museum operations depend heavily on visitor behavior analysis, which serves as essential input for both strategic planning and everyday activities. Digital tracking technologies exceed conventional survey and manual observation methods by collecting visitor information, including movement patterns and dwell times, along with interaction sequences and social dynamics. Exhibition design choices, along with content curation and resource allocation decisions, receive support through this analytical process. Smart museums need to understand visitor behavior to improve the personalized experiences they provide. The combination of advanced technology with visitor interaction at these heritage sites leads to distinctive engagement opportunities [13]. Smart museum design now requires the customization of content and experiences based on unique visitor profiles [14].

The study of museum experiences needs a multidisciplinary method which integrates technological tools with an understanding of human behavior [15]. Visitors build their understanding through cognitive and emotional activities instead of just absorbing information. The design of personalized technologies requires thorough knowledge of perception, learning, and engagement because these elements are crucial to the quality of visitor experiences. Museum experience enhancement depends on understanding how emotions and cognitive processes interact during visitor engagement. Research indicates that the design of interactive technologies must incorporate emotional and affective elements to establish significant connections with cultural content. For instance, Fominska et al. [16] advocated the application of modern digital tools alongside cognitive-emotional theories to detect and study the emotional reactions of museum visitors. Through this strategy, museums can establish responsive systems, which personalize the learning experience by adapting to individual visitor requirements.

Modern research shows progress in museum digitization, though it lacks sufficient insight into how visitor behavior analysis can improve personalized experiences. Multiple technologies have been well-documented according to studies by Li et al. [10] and Furferi et al. [17]. Despite technological advancements in museum digitization, no research thoroughly investigates both visitor behavior analytics and personalization strategies. Current research [18,19] centers around technology application or visitor experience studies yet fails to connect these fields to develop evidence-based, behavior-focused personalization strategies.

The study by Nigatu et al. [20] reveals exponential growth in museum visitor experience research since 2017, though it also points out geographical research disparities with insufficient representation from varied cultures. There exists a requirement for a systematic review that combines worldwide perspectives and uncovers culturally appropriate methods for analyzing visitor behavior.

Multiple systems now operate simultaneously within museum technologies, creating a complex ecosystem explained through Lu et al.’s four-stage evolutionary model [15]. The Internet of Cultural Things framework developed by Liang et al. [21] shows how integrated systems are changing traditional collections into dynamic environments. We still lack adequate knowledge about the best methods to integrate these emerging technologies so they can effectively engage with visitor interactions.

In earlier research, Kosmopoulos and Styliaras [22] presented an analysis of architectures for personalized content delivery services, which, while valuable, could not anticipate current challenges in real-time behavior analytics due to its publication timeframe. Studies consistently identify implementation challenges that hinder the effective deployment of personalization systems in museums. These include the “cold start” problem in profile creation, interface usability across demographic groups [23], and balancing technological innovation with cultural authenticity [19].

Research into visitor behavior analysis shows significant methodological diversity, with approaches ranging from technology acceptance models [23] to inclusive design principles [24]. Experts have yet to reach an agreement about which methodological approaches work best for various visitor segments and institutional settings. A systematic review would compare these various methods and determine the best practices for different scenarios.

The study by Kasemsarn et al. [24] emphasized the critical need to enhance museum accessibility for older adults and individuals with mobility impairments. However, there remains limited understanding of how visitor behavior analysis can be leveraged to create more inclusive experiences across diverse user groups. The current understanding of how visitor behavior analysis can be applied to develop inclusive experiences for diverse user groups remains minimal. The growing practice of collecting and analyzing visitor data introduces serious ethical and privacy issues which existing research does not fully explore. The study by Li et al. [10] discusses digital transformation challenges but lacks thorough exploration of ethical frameworks for visitor behavior analysis.

Li et al. [10] documented the accelerated deep transformation of the museum sector resulting from the COVID-19 pandemic. Since then, institutions have continued to discover the advantages of utilizing data-driven approaches for decision-making. There is currently no strategic framework that addresses how visitor behavior analysis will evolve to meet future challenges. A systematic review would uncover emerging patterns and predict future progress, which would form the basis of research direction and priority recommendations.

Although personalization offers significant advantages, such as better engagement and learning results along with visitor satisfaction, it prompts important discussions about privacy issues, accessibility challenges, and the essential purpose of museums in society.

This review examines how to combine existing knowledge with theoretical principles to create a research-supported framework that guides museums when they implement visitor behavior analysis systems.

1.2. Conceptual Framework for Visitor Behavior, Technology, and Outcomes

Understanding how visitors’ behavior links with technological innovation and desired outcomes is essential for managing modern digitized and personalized museum experiences. This conceptual framework systematically represents the interrelationships between behavioral parameters and technological solutions leading to target outcomes within an integrated model.

Three components form the core structure of the framework shown in Figure 1. The first component examines visitor behavioral parameters that cover cognitive functions alongside emotional responses and social and physical interactions within the museum environment. These parameters play a crucial role in uncovering the intricate interactions between museum visitors and exhibits. Attention span, time spent with exhibits, and social behaviors provide insights into both personal choices and the dynamics of group activities. The conceptual foundation of this element aligns with Falk and Dierking’s contextual learning model [3], which highlights learning impacts through personal, social, and physical dimensions.

The second component of the framework is technological solutions that act as intermediaries between museum institutions and visitors. The technologies range from artificial intelligence (AI) and machine learning (ML) systems to location tracking tools as well as virtual and augmented reality (VR/AR) technologies and interactive display equipment. These technological innovations function to gather visitor behavior data while enabling adaptive content personalization opportunities.

The third component consists of desired outcomes which integrate the cognitive, emotional, and social objectives of a museum visit. These outcomes consist of better content for visitors along with more visitor satisfaction and enhanced social involvement. Smart technologies that create personalized experiences can fulfill visitors’ basic psychological needs and enhance the satisfaction level of museum exhibit interactions.

Technologies modify content based on visitor behavior data through an iterative interaction process, while resulting feedback helps improve future technological solutions. A dynamic relationship represents the intricate nature of modern museum experiences through continuous exchanges among data sources and technology systems alongside visitors’ requirements.

The proposed conceptual framework functions as both an analytical device to study smart museum interactions and a practical manual for creating successful personalization methods.

1.3. Research Questions

This systematic review addresses three interrelated research questions that emerge from critical gaps in smart museum literature and practice. These questions form a comprehensive framework for understanding the complex relationship between visitor behavior analysis, technological implementation, and experiential outcomes in smart museum environments.

The research questions guiding this systematic review are as follows:

• RQ1: What are the current methodologies for analyzing visitor behavior in smart museums?
• RQ2: Which technologies and the analysis of what museum visitor behaviors contribute to content personalization?
• RQ3: Which personalization technologies and strategies are most used and promising?

The first research question assesses various methods used to analyze visitor behavior in technologically enhanced museums. The study investigates traditional observational methods alongside new digital approaches, with special consideration given to the technical infrastructure that supports these methods. The second research question explores the way technology and visitor behavior analysis work together to produce customized museum experiences. The research examines exhibit dwell time together with digital interaction analysis and feedback systems while monitoring revisitation patterns and visitor demographic preferences. The third research question examines which personalization technologies and approaches become most common and successful in smart museums.

This systematic review seeks to merge current knowledge and pinpoint research gaps that need more investigation. Research results from these questions will enhance both theoretical knowledge and practical use in the field of smart museums while supporting the continuous development of personalized cultural experiences alongside technological integration within museum settings.

1.4. Review Scope and Objectives

The review period spans ten years, from January 2015 to January 2025, and covers numerous significant advances in museum digitization technology. This chronological framework captures several critical evolutionary phases: museums adopted Internet of Things (IoT) technology; artificial intelligence and machine learning expanded to analyze visitor data; digital transformation sped up due to global socioeconomic changes; and the sector accumulated extensive longitudinal data to spot trends and patterns.

Multiple analytical dimensions serve as the framework for examining visitor behavior parameters. Physical parameters consist of visitor movement and navigation routes through exhibits, time spent at displays, spatial heat maps showing concentration areas, patterns of visitor traffic flow, and how groups behave together. Museum digital interaction parameters feature exhibit engagement measurements while also analyzing mobile application usage patterns together with interactive display interaction times and digital content consumption habits along with social media interactions within museum environments. Cognitive and emotional parameters research the patterns of attention distribution, along with learning behavior displays and exhibit-induced emotional responses and decision-making mechanisms, as well as information retention signals.

The examination of personalization aspects takes place across various conceptual frameworks. Content personalization systems integrate adaptive delivery methods together with linguistic and accessibility adaptations, as well as cultural context considerations along with learning style accommodations and interest-based filtering systems. Experience customization research includes analysis of tour route optimization algorithms together with interactive exhibit adaptation systems and real-time feedback integration protocols, as well as preference learning mechanisms and visit-duration optimization frameworks.

The review strives to integrate current knowledge about visitor behavior analysis methods in smart museums while assessing diverse personalization techniques for effectiveness and identifying best practices for smart museum technology implementation.

2. Methodology

2.1. Search Strategy

The systematic review process successfully found literature that studied smart museum technologies and personalized visitor experiences. The research team picked databases to ensure coverage of museum studies alongside computer science, information systems, and cultural heritage studies. Three primary academic repositories were utilized: The Web of Science Core Collection, Scopus, and IEEE Xplore.

The search strategy emphasizes building search strings which are centered on critical conceptual areas pertinent to our research. The research areas included smart museum contexts described by terms like “smart museums”, “digital museums”, “interactive museums”, “intelligent museums”, and “personalized museums”; visitor behavior analysis, which involved terms such as “visitor behavior”, “audience engagement”, “user interaction”, “visitor experience”, “visitor mobility”, “visitor movement”, “visitor tracking”, and “visitor patterns”; and technological innovations, which covered “technologies”, “innovations”, “artificial intelligence”, “AI”, and “machine learning”. The search strings implemented for Scopus, IEEE Xplore, and Web of Science databases appear in

Table 1. Search strings.

Database	Search String
Scopus	(“smart museums” OR “digital museums” OR “interactive museums” OR “intelligent museums” OR “personalized museums”) AND (“visitors behavior” OR “audience engagement” OR “user interaction” OR “visitor experience” OR “visitor mobility” OR “personalized content delivery”) AND (“technology” OR “technologies” OR “innovations” OR “artificial intelligence”) (TITLE (“museum*” AND “behavi*”)) AND (LIMIT-TO (SRCTYPE, “j”)) (TITLE (“museum*” AND “personal*”)) AND (LIMIT-TO (SRCTYPE, “j”))
IEEE Xplore	((“museum*” OR “cultural heritage” OR “exhibition*”) AND (“visitor behavio*” OR “visitor analytic*” OR “visitor track*” OR “visitor* movement” OR “visitor mobility*” OR “visitor pattern*”))
WoS	TS = (“smart museums” OR “digital museums” OR “interactive museums” OR “intelligent museums” OR “personalized museums”) AND TS = (“visitors behavior” OR “audience engagement” OR “user interaction” OR “visitor experience” OR “personalized content delivery”) AND TS = (“technology” OR “technologies” OR “innovations” OR “artificial intelligence” OR “AI” OR “machine learning”)

.

Detailed results for the retrieved records when applying the selected search strings to the three databases are shown in

Table 2. Number of unique publications.

Database	Records After Search	Invalid Records	Duplicates Within Database	Duplicates Between Databases	Total Records
Scopus	1882	132	573	-	1177
IEEE Xplore	55	0	0	8	47
Web of Science	5	0	0	4	1
Total	1942	132	573	12	1225

.

The search parameters were narrowed to target only publications in English from 2015 to the end of December 2024, including articles and conference papers. The systematic review required specific inclusion and exclusion criteria, which allowed researchers to choose relevant studies of high quality (see

Table 3. Criteria used for including and excluding research studies.

Inclusion Criteria	Exclusion Criteria
1. Peer-reviewed scientific publications presenting innovative smart museum implementations (journal articles and conference papers) 2. Evidence of visitor behavior analysis or personalization strategies 3. Relevant to research questions (research focus) 4. English-language publications 5. Published between 2015 and the present (end of December 2024)	1. Non-English version 2. Lack of abstract 3. No access permission 4. Publication not in a journal or conference 5. Content does not match the research questions 6. Published before 2015

). The inclusion criteria selected peer-reviewed scientific publications, including journal articles and conference papers which demonstrated new smart museum implementations. The research needed to demonstrate visitor behavior analysis or show methods for tailoring museum experiences. Studies needed to match the review’s specific research focus and questions to be selected.

Research that failed to meet the required standards or relevance was removed through the application of exclusion criteria. The review excluded studies published in non-English languages along with those without an abstract because abstracts are needed for initial screening and relevance assessment. Research inaccessible through permission or subscription barriers was eliminated from consideration. The review did not include publications that were not featured in established academic journals or conference proceedings. Any research that failed to address the review’s specific questions was removed from consideration along with publications from before 2015.

2.2. Screening Process

The initial systematic search of Scopus, IEEE Xplore, and Web of Science produced 1942 records. According to the Table 2 results, Scopus provided the largest record count (n = 1882), then IEEE Xplore showed fewer records (n = 55), and Web of Science returned the least number of records (n = 5). The research team executed a two-step deduplication method prior to formal screening, which eliminated 573 intra-database duplicates and 12 inter-database duplicates. The initial screening phase included 1225 unique publications after identifying and removing the 132 records deemed invalid.

2.2.1. Screening Process Methodology

We offer a new methodology executed by Python software (version 3.9.13) applications to tackle the lengthy publication screening process in systematic reviews (Supplementary Material/Software). The proposed methodology achieves faster screening times without compromising on methodological integrity or thematic consistency. A multi-stage computational method served as the basis for the screening process. The initial stage of retrieval involved programmatically accessing three scientific databases through their Application Programming Interfaces (APIs) with predefined search strings to obtain potentially relevant publications. After collection from databases, the corpus received initial filtering to remove publications that failed to meet inclusion standards, including non-journal and non-conference sources plus those published before 2015. The system removed duplicate entries by using automated cross-reference verification processes. The remaining publications underwent semantic similarity analysis through the Sentence-BERT (S-BERT) framework developed by Reimers and Gurevych [25] by applying the paraphrase-MPNet-base-v2 model. Combining Siamese BERT-Network elements with MPNet’s masked and permuted pre-training yields advanced contextual embeddings which discern semantic differences beyond basic lexical matches [26].

The model produced high-dimensional vector representations of review themes alongside content signatures from publication titles, abstracts, and keywords, then calculated cosine similarity scores to determine thematic relevance percentages. The ranked corpus includes bibliographic details along with abstracts and keywords while also providing document typology and citation metrics together with percentages showing thematic similarity levels (see Figure 2).

Semantically relevant terms related to the focus of the review were programmatically colored with different background colors to enable reviewers to visually identify pertinent content during manual publication screening. The screening method developed for this research project successfully preserved the corpus’s relevance to the study objectives.

2.2.2. PRISMA Flowchart

The study selection process followed PRISMA [27] guidelines, which provided systematic reviews with methodological rigor and transparency (Supplementary Table S1). The protocol for screening included explicit inclusion and exclusion criteria, as shown in Figure 3. The publications that did not match current smart museum practices were identified and removed from consideration by our proposed methodology based on their publication dates, totaling 278 works. During the title and abstract screening stage, we excluded 896 publications because they failed to align with our research targets, which studied visitor behavior analysis and personalized experiences in smart museum settings.

We found 51 studies that needed full-text reports for further comprehensive evaluation after performing our initial filtering procedures. The inability to acquire two reports necessitated their exclusion from subsequent assessments. The predefined eligibility assessment excluded 16 publications from further analysis since their content did not sufficiently address visitor behavior analytics and personalization methods. The research included 33 studies that met all selection criteria for final analysis. For our qualitative synthesis, we selected these publications as the core dataset because they provided fundamental insights into visitor behavior analysis and personalization techniques within smart museums.

Table 4. List of selected publications.

Doc ID	Document Title	Document Type	Authors	Year
1	Implicit engagement detection for interactive museums using brain-computer interfaces	Conference paper	Abdelrahman et al. [28]	2015
2	Using personas to model museum visitors	Conference paper	Almeshari et al. [29]	2019
3	Museum mobile guide preferences of different visitor personas	Journal paper	Almeshari et al. [30]	2021
4	Measuring and evaluating visitors’ behaviors inside museums: the Co. ME. project	Journal paper	Angeloni et al. [31]	2021
5	Pepper4Museum: towards a human-like museum guide	Conference paper	Castellano et al. [32]	2020
6	Early prediction of visitor engagement in science museums with multimodal learning analytics	Conference paper	Emerson et al. [33]	2021
7	Using deep learning for collecting data about museum visitor behavior	Journal paper	Ferrato et al. [34]	2022
8	The META4RS Proposal: museum emotion and tracking analysis for recommender systems	Conference paper	Ferrato et al. [35]	2022
9	Museum visitor activity tracker using indoor positioning system	Conference paper	Handojo et al. [36]	2019
10	Exploiting behavioral user models for point of interest recommendation in smart museums	Journal paper	Hashemi and Kamps [37]	2018
11	Exhibitxplorer: Enabling personalized content delivery in museums using contextual geofencing and artificial intelligence	Journal paper	Ivanov [38]	2023
12	Advanced Visitor Profiling for Personalized Museum Experiences Using Telemetry-Driven Smart Badges	Journal paper	Ivanov [39]	2024
13	Delivering Personalized Content to Open-air Museum Visitors Using Geofencing	Conference paper	Ivanov and Velkova [40]	2022
14	Personalizing cultural heritage access in a virtual reality exhibition: A user study on viewing behavior and content preferences	Conference paper	Javdani et al. [41]	2023
15	SARIM: A gesture-based sound augmented reality interface for visiting museums	Conference paper	Kaghat et al. [42]	2018
16	Personalized multimedia content delivery on an interactive table by passive observation of museum visitors	Journal paper	Karaman et al. [43]	2016
17	The Impact of Immersive Technology in Museums on Visitors’ Behavioral Intention	Journal paper	Liu and Sutunyarak [44]	2024
18	Exploiting density to track human behavior in crowded environments	Journal paper	Martella et al. [45]	2017
19	Leveraging proximity sensing to mine the behavior of museum visitors	Conference paper	Martella et al. [46]	2016
20	Visualizing, clustering, and predicting the behavior of museum visitors	Journal paper	Martella et al. [47]	2017
21	RECITE: A framework for user trajectory analysis in cultural sites	Journal paper	Orenes-Vera et al. [48]	2021
22	SeSAME: Re-identification-based ambient intelligence system for museum environment	Journal paper	Paolanti et al. [49]	2022
23	Exploring the potential of the internet of things at a heritage site through co-design practice	Conference paper	Petrelli et al. [50]	2018
24	A Holistic Approach for Enhancing Museum Performance and Visitor Experience	Journal paper	Philippopoulos et al. [51]	2024
25	An evolving museum metaphor applied to cultural heritage for personalized content delivery	Journal paper	Rajaonarivo et al. [52]	2019
26	Tracking visitors in a real museum for behavioral analysis	Conference paper	Rashed et al. [53]	2016
27	Build your own hercules: Helping visitors personalize their museum experience	Conference paper	Rey et al. [54]	2020
28	Visiting virtual museums: How personality and art-related individual differences shape visitor behavior in an online virtual gallery	Journal paper	Rodriguez-Boerwinkle and Silvia [55]	2024
29	Flow, staging, wayfinding, personalization: Evaluating user experience with mobile museum narratives	Journal paper	Roussou and Katifori [56]	2018
30	RECBOT: Virtual Museum navigation through a Chatbot assistant and personalized Recommendations	Conference paper	Tsitseklis et al. [57]	2023
31	Personalization in digital ecomuseums: The case of Pros-Eleusis	Journal paper	Vrettakis et al. [58]	2023
32	Analysis of visitors’ mobility patterns through random walk in the Louvre Museum	Journal paper	Yoshimura et al. [59]	2024
33	What influences user continuous intention of digital museum: integrating task-technology fit (TTF) and unified theory of acceptance and usage of technology (UTAUT) models	Journal paper	Zheng et al. [60]	2024

shows detailed data about each document, including its Doc ID, title, publication type, author details, and publication year.

2.3. Quality Assessment Criteria

We ensured our results’ reliability and validity through systematic evaluations of selected studies’ methodological rigor. Our systematic evaluation of methodological rigor involved utilizing the QualSyst framework developed by Kmet et al. [61] that employs a flexible evaluation system compatible with both quantitative and qualitative methods (see

Table 5. Criteria for assessing the methodological rigour of selected studies.

Criteria	Sub-Criteria	Scoring, Points
Methodological Clarity	Description of research design, appropriateness, detailed documentation, replicability	0–8
Data Collection and Analysis	Sample size, data collection procedures, reliability of techniques, statistical methods	0–8
Technical Performance	Description of technological solutions, system architecture, performance metrics, implementation details	0–8
Reporting of Results	Presentation of findings, supporting evidence, use of tables/figures, discussion of limitations	0–8

).

The evaluation framework reviews four essential domains, including methodology clarity together with data collection techniques and analysis methods, performance standards, and how results are reported. A maximum of 8 points was assigned to each criterion, while every sub-criterion could earn scores between 0 and 2 points depending on the fulfillment level (0 = unmet criterion, 1 = partially met criterion, 2 = fully met criterion). Three quality classifications emerged from the scoring system: high-quality studies scored between 24 and 32 points, while average-quality studies scored between 16 and 23 points, and studies of low quality received scores between 0 and 15 points.

Figure 4 summarizes the reviewed articles’ mean performance in four distinct criteria. The chart shows that Methodological Clarity achieved the highest score with an average of 7.82 points, which represents 97.73% of the full score potential. The research design and procedures were clearly defined in most articles, which ensured they could be replicated and that they maintained methodological rigor. The Data Collection and Analysis criterion received a mean score of 7.48, which translates to 93.56% of the maximum score. The performance demonstrated strength in sample size justification alongside data reliability and appropriate analytical techniques. Technical Performance achieved a mean score of 7.09, or 88.64%, slightly lower than the preceding criteria. This indicates that while many studies provided adequate details about technological solutions, performance metrics, or system architecture, some fell short in offering a comprehensive description. Reporting results were scored the lowest among the individual criteria, with a mean of 6.97, representing 87.12% of the maximum. This result highlights recurring issues in presenting findings clearly, using appropriate evidence, and adequately discussing limitations. The combined score from all four criteria totals 29.36, which represents 91.76% of the highest achievable score of 32.

3. Results

3.1. Publication Trends and Bibliometric Analysis

3.1.1. Temporal Distribution of Publications

Figure 5 illustrates that research interest in visitor behavior and personalized experiences in smart museums has shown a distinct upward trend over time. During the years 2021 through 2024, researchers produced 18 publications, which make up 54.5% of the total publications in this area, showing an increase in research activity recently. The observed research trend demonstrates increasing recognition of visitor behavior analysis and personalization importance within museums, which aligns with technological advancements in IoT, AR, and intelligent systems.

3.1.2. Geographical Distribution of Publications

Figure 6 shows that European countries host most research examining smart museums and visitor behavior. The leading country in terms of publications about smart museums and visitor behavior is Italy, with seven contributions representing 17.1%, while the Netherlands follows with five publications accounting for 12.2%. Four publications each were submitted by the United States, France, and Greece, making up 9.8% of the total publications from each country. The cultural wealth of Europe, along with its established museum systems and the European Digital Agenda funding, contributes to the continent’s dominant role. Research spanning 17 countries across five continents reveals that improving visitor experiences in cultural institutions holds global importance but shows different levels of research focus.

3.1.3. Publication Types

A review of publication types (Figure 7) revealed that journal articles made up the largest segment with 19 publications (57.6%), while conference papers comprised 14 publications (42.4%). The extensive number of journal articles reveals a field with mature methodologies and theoretical bases, while the substantial quantity of conference papers demonstrates the presence of active technical developments and real-world applications.

3.1.4. Thematic Analysis

The analyzed literature’s principal themes and concepts are displayed through a word cloud visualization in Figure 8. This literature examines main thematic areas, which include “Museums”, “Cultural heritages”, “Behavioral research”, and “Augmented reality”. This field shows its interdisciplinary nature by combining technological terms such as “Internet of things”, “Intelligent systems”, and “Augmented reality” with visitor-focused concepts like “Museum visitor”, “Personalization”, and “Exhibitions”.

3.2. Methodologies

3.2.1. Selected Core Methodologies

Table 6. Overview of methodology segments.

Methodology Segment	Abbreviation	Number of Papers (%) [Doc. IDs]	Analytical Significance
Statistical and Data Analysis	SDA	32 (97.0) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33]	Foundational quantitative approach
AI and Machine Learning Approaches	AML	21 (63.6) [2, 4, 5, 6, 7, 8, 10, 11, 12, 14, 16, 18, 19, 20, 21, 22, 24, 25, 26, 30, 31]	Advanced predictive analytics
Mobile and Interactive Technologies	MIT	20 (60.6) [2, 3, 5, 6, 8, 9, 10, 11, 12, 13, 15, 16, 19, 23, 24, 25, 27, 29, 30, 31]	Technological interaction mapping
Geospatial Methods	GSM	15 (45.5) [5, 7, 9, 10, 11, 12, 13, 15, 18, 20, 21, 25, 26, 28, 32]	Contextual spatial understanding
Survey and Interview-Based Methods	SIB	13 (39.4) [1, 2, 3, 12, 14, 15, 17, 23, 27, 28, 29, 31, 33]	Qualitative behavioral observations
Wireless Signal-Based Tracking	WST	11 (33.3) [9, 11, 12, 13, 18, 20, 21, 23, 24, 27, 31, 32]	Spatial movement reconstruction
Computer Vision and Sensor-Based Methods	CVS	10 (30.3) [4, 5, 6, 7, 8, 14, 15, 16, 22, 26]	Non-invasive behavioral capture
Biometric and Physiological Monitoring	BPM	5 (15.2) [1, 5, 6, 8, 14]	Intrinsic emotional response tracking
Virtual/Augmented Reality Methods	VAR	4 (12.1) [14, 15, 25, 28]	Immersive experience analysis

presents details about the methodological segments used in this review, including their abbreviations and the number of papers using each methodology, along with paper IDs (Doc. IDs), their percentage distribution, and their analytical significance.

The field is predominantly led by Statistical and Data Analysis, which accounts for 97% and serves as the primary foundation for quantitative methods. The adoption rate for AI and Machine Learning stands at 63.6%, while Mobile and Interactive Technologies reach 60.6%, which underscores the significance of predictive analytics and technological interaction mapping. Survey and Interview-Based Methods (39.4%), along with Geospatial Methods (45.5%), achieve moderate implementation by merging spatial data analysis with qualitative research findings. Wireless Signal-Based Tracking holds a 33.3% usage rate, while Computer Vision and Sensor-Based Methods appear 30.3% of the time and Biometric and Physiological Monitoring registers at 15.2%, with Virtual/Augmented Reality Methods coming in at 12.1%.

This distribution demonstrates a methodological progression from basic statistical analysis toward advanced technological approaches, with emphasis on mobile technologies and spatial analytics for understanding visitor behavior in smart museum environments.

3.2.2. Methodological Complexity

A critical insight from our analysis is the increasing sophistication of methodological approaches. The distribution of methodological complexity reveals a significant trend towards multi-methodology research designs:

• Papers using two methodologies:…………… 4 (12.1%)
• Papers using three methodologies: ……………7 (21.2%)
• Papers using four methodologies: ……………13 (39.4%)
• Papers using five methodologies:…………… 5 (15.2%)
• Papers using six methodologies:…………… 3 (9.10%)
• Papers using seven methodologies: ……………1 (3.00%)

The dataset reveals that researchers show a strong preference for multi-methodology research designs, which is demonstrated by 39.4% of papers using four distinct methodologies. The fact that only 12.1% of papers included two methodologies demonstrates a consensus among researchers regarding this approach’s insufficiency to capture visitors’ behavior complexity in smart museum settings. Research designs that combine 6–7 methodologies are uncommon (12.1% combined) because of implementation challenges and resource limitations, which make them difficult to execute and process alongside diminishing analytical benefits.

The bell-shaped distribution centered around four methodologies. The movement toward combined research methods matches smart museum research, which requires simultaneous capture of technological, spatial, psychological, and social aspects to study visitor behavior.

3.2.3. Methodology Combinations

When researchers combine methodologies, they uncover a complex research environment that shows that specific dyadic and multi-method approaches dominate.

Table 7. Distribution of top methodology combinations.

Methodology Combination	Number of Papers (%)	Doc IDs
AML + SDA	21 (63.6)	[2, 4, 5, 6, 7, 8, 10, 11, 12, 14, 16, 18, 19, 20, 21, 22, 24, 25, 26, 30, 31]
MIT + SDA	19 (57.6)	[2, 3, 5, 6, 8, 9, 10, 11, 12, 13, 15, 16, 19, 23, 24, 25, 29, 30, 31]
GSM + SDA	15 (45.5)	[5, 7, 9, 10, 11, 12, 13, 15, 18, 20, 21, 25, 26, 28, 32]
AML + MIT + SDA	13 (39.4)	[2, 5, 6, 8, 10, 11, 12, 16, 19, 24, 25, 30, 31]
AML + GSM + SDA	10 (30.3)	[5, 7, 10, 11, 12, 18, 20, 21, 25, 26]
AML + CVS + SDA	9 (27.3)	[4, 5, 6, 7, 8, 14, 16, 22, 26]
GSM + MIT + SDA + WST	6 (18.2)	[9, 10, 11, 12, 13, 15]
AML + GSM + SDA + WST	6 (18.2)	[10, 11, 12, 18, 20, 21]
AML + GSM + MIT + SDA	5 (15.2)	[5, 10, 11, 12, 25]

shows how the top methodology combinations are distributed, which helps understand the most used analytical strategies.

The most frequent dual focus of research involves AI and Machine Learning together with Statistical and Data Analysis (AML + SDA), which appears in 21 scientific papers. Advanced predictive analytics normally extends fundamental statistical approaches, which serve as complementary methodologies. Research on technological interactions in smart museums shows the importance of quantitative analysis through 19 papers on Mobile and Interactive Technologies combined with Statistical and Data Analysis (MIT + SDA). Geospatial Methods and Statistical and Data Analysis (GSM + SDA) feature in 15 research articles, which illustrates how quantitative methods are effective in examining spatial visitor behavior patterns. Statistical and Data Analysis (SDA) features in every top combination, which proves its fundamental importance in visitor behavior research. The regular application of data-driven methods reveals their essential role in creating and improving individualized visitor experiences in smart museums. Research shows that 13 different studies combine AI, machine learning methods with mobile technologies, and statistical data analysis to enable deep investigation of visitor behavior patterns. The application of four-method combinations, including GSM + MIT + SDA + WST alongside AML + GSM + SDA + WST, represents an effort to merge various technological tools with analysis instruments to achieve advanced visitor interaction studies.

3.2.4. Meta-Analysis

Statistical and Data Analysis (SDA) holds the highest adoption rate at 97.0% while remaining well ahead of AI and Machine Learning (AML) at 63.6% with a Z score of 3.41 (p < 0.001) and Mobile and Interactive Technologies (MIT) at 60.6% with a Z score of 3.62 (p < 0.001). The combination of AML and SDA represents the most prevalent methodological pairing in 63.6% of the papers analyzed (21/33), while MIT and SDA follow at 57.6% (19/33), and both show significance above a random distribution according to Fisher’s exact test with p-values below 0.01. The specialized research methods Biometric and Physiological Monitoring (BPM) and Virtual/Augmented Reality Methods (VAR) show significantly lower adoption rates at 15.2% and 12.1%, respectively.

3.3. Technologies

3.3.1. Selected Core Technologies

The digital transformation of museum spaces has catalyzed the integration of diverse technological solutions aimed at enhancing visitor experiences through personalization and behavioral analysis. Our systematic review identified twelve core technological categories instrumental in facilitating smart museum environments (see

Table 8. Selected core technologies.

Core Technology	List of Sub-Technologies
AI/ML Systems	Recommender systems, personalization engines, chatbots
Behavioral Analytics Platforms	Visitor analytics, heatmap generation, path analysis
Biometric Sensors	Eye tracking, EEG devices, wearable sensors, heart rate monitors
Computer Vision Systems	RGB cameras, depth sensors, facial detection, LIDAR, person tracking
Content Management Systems	Digital asset management, personalized content delivery
Interactive Displays	Touchscreens, interactive tables, digital signage
IoT Sensors	Proximity sensors, environmental sensors, smart objects
Location Tracking Systems	BLE beacons and badges, RFID/NFC tags, GPS, Wi-Fi positioning, IR sensors, geofencing
Mobile and Wearable Devices	Smartphones, tablets, audio guides, wearable devices
Social Media Integration	Social network analysis, sharing platforms, community features
Survey/Feedback Systems	Digital questionnaires, interactive feedback tools
Virtual/Augmented Reality	VR headsets, AR applications, mixed reality systems

).

The current technology environment includes traditional systems like Content Management Systems and Mobile and Wearable Devices, as well as new developments such as Biometric Sensors and Virtual/Augmented Reality applications. AI/ML Systems fulfill a critical analytical role as recommender systems and personalization engines convert behavioral data into actionable insights. Behavioral Analytics Platforms enable this transformation by processing visitor movement patterns along with engagement metrics and interaction preferences.

Computer Vision Systems and IoT Sensors with Location Tracking technologies make up the physical monitoring infrastructure that spatially maps visitor behaviors with precise temporal and positional accuracy. The systems enable analysis of crowd behavior at a macro scale while tracking individual visitor preferences at a micro level without requiring visitor inputs. Explicit visitor sentiment data from Survey/Feedback Systems and Social Media Integration serve as contextual information for behavioral metrics gathered implicitly through the museum’s tracking systems.

Interactive Displays, Mobile and Wearable Devices, and Virtual/Augmented Reality systems function as experiential interfaces that link visitors to museum content while collecting data and delivering personalized content. The complex interactions among these technological elements lead to responsive museum spaces that adjust based on individual visitor traits.

3.3.2. Analysis of Base Technologies

The review of 33 publications shows how core technologies distribute and appear to create personalized experiences in smart museums.

Table 9. Core technologies used.

Technology	Number of Papers (%)	Doc IDs
AI/ML Systems	11 (33.3)	[1, 4, 5, 8, 10, 11, 12, 20, 22, 25, 30]
Behavioral Analytics Platforms	14 (42.4)	[7, 8, 9, 12, 18, 19, 20, 21, 22, 25, 26, 28, 29, 32]
Biometric Sensors	5 (15.2)	[1, 6, 8, 14, 28]
Computer Vision Systems	9 (27.3)	[4, 5, 6, 7, 8, 16, 22, 26, 28]
Content Management Systems	9 (27.3)	[2, 11, 12, 13, 16, 21, 24, 25, 30]
Interactive Displays	5 (15.2)	[6, 16, 17, 23, 29]
IoT Sensors	9 (27.3)	[4, 5, 10, 12, 18, 19, 23, 27, 32]
Location Tracking Systems	13 (39.4)	[9, 10, 11, 12, 13, 15, 18, 19, 20, 21, 24, 31, 32]
Mobile and Wearable Devices	14 (42.4)	[1, 2, 9, 10, 11, 12, 13, 15, 16, 21, 24, 29, 31, 32]
Social Media Integration	2 (6.0)	[27, 33]
Survey/Feedback Systems	11 (33.3)	[2, 3, 11, 12, 13, 14, 17, 27, 28, 29, 31]
Virtual/Augmented Reality	5 (15.2)	[14, 15, 17, 25, 28]

shows a listing of technological implementations in the studies and includes the quantity of papers and document IDs for each technology.

Most Frequently Used Technologies

Fourteen papers utilized Behavioral Analytics Platforms and Mobile and Wearable Devices as the top technologies, with both appearing in 42.4% of cases. The data demonstrate a dedicated approach to studying visitor actions while using mobile technology and wearables for customized experience delivery. The use of Location Tracking Systems in 13 papers (39.4%) demonstrates their importance for tracking museum visitor movements to customize content and experiences. Both AI/ML Systems and Survey/Feedback Systems show substantial use across 11 papers (33.3%), indicating museums heavily depend on artificial intelligence to create personalized experiences while using visitor feedback to enhance museum services.

Moderately Used Technologies

Nine publications use Computer Vision Systems, Content Management Systems, and IoT Sensors, which represent 27.3% of the research papers. The integration of these technologies enables museums to analyze visitor behavior while managing digital assets and collecting environmental information to improve the visitor experience. Biometric Sensors, Interactive Displays, and Virtual/Augmented Reality technology all appeared in five research papers, which account for 15.2% of the cases. The reduced usage of these technologies suggests they represent emerging or specialized applications in smart museums despite their innovative visitor engagement potential.

Least Used Technology

Social Media Integration is the least used technology, appearing in only two papers (6.0%). This indicates that while social media may play a role in visitor engagement, it is not yet a widely adopted tool for personalization in smart museums. This presents a challenge, as recent experiments conducted by Alqurafi and Alsanoosy [62] demonstrated that machine learning methods yield significantly improved results (with an accuracy of up to 93.3% for the LSTM model in sentiment analysis based on online reviews).

3.3.3. Meta-Analysis

A meta-analysis of the data reveals distinct patterns in technology adoption across smart museums. Descriptive trends show marked variation. Behavioral Analytics Platforms and Mobile/Wearable Devices (each 42.4%, n = 14) exhibit Z-scores of +1.70 relative to the mean adoption rate (μ = 8.92 mentions per technology, σ = 2.99), indicating elevated usage. Social Media Integration (6.0%, n = 2) has a Z-score of −2.31, confirming its outlier status. Moderately adopted technologies like Computer Vision Systems (27.3%, n = 9) and IoT Sensors (27.3%, n = 9) align near expectation (Z = +0.03), while niche technologies such as Biometric Sensors (15.2%, n = 5) and Virtual/Augmented Reality (15.2%, n = 5) fall below (Z = −1.31).

3.4. Visitor Behavior

3.4.1. Data Collection Methods

The ability to deliver personalized museum content depends heavily on effective data collection methods. Analysis of visitor behavior and preferences through data collection helps create improved user experiences. The publications studied in this research use a combination of questionnaires, computer vision, and other techniques, including BLE beacons and RFID/NFC, to gather data. The primary categories of data collection methods, along with their descriptions and publication identifiers for each category, are presented in

Table 10. Data collection method group names.

Group Name	Description	Paper IDs
Brain–Computer Interface (BCI)	Data collected using EEG signals and brain–computer interfaces to detect cognitive and emotional states.	1
Questionnaires/Surveys	Data collected through face-to-face or online questionnaires and surveys to gather visitor preferences, demographics, and feedback.	2, 3, 14, 17, 27, 28, 33
Computer Vision	Data collected using cameras and computer vision techniques to track visitor movements, facial expressions, and interactions with exhibits.	4, 5, 7, 8, 16, 22
Bluetooth Low Energy	Data collected using BLE beacons and sensors to track visitor locations and movements within the museum.	9, 11, 12, 13, 21, 24, 32
RFID/NFC	Data collected using RFID tags or NFC technology to track visitor interactions with exhibits and provide personalized content.	10, 23, 27
Proximity Sensors	Data collected using proximity sensors to track visitor movements, density, and interactions with exhibits.	18, 19, 20
Multimodal Sensors	Data collected using multiple sensors (e.g., eye gaze, facial expression, and posture tracking) to analyze visitor behavior.	6
GPS/Geofencing	Data collected using GPS and geofencing technologies to track visitor locations and provide location-based content.	11, 13, 24, 31
LIDAR	Data collected using LIDAR sensors to track visitor positions and movement patterns.	26
Virtual Reality/Augmented Reality (VR/AR)	Data collected through VR environments and interactions to analyze visitor behavior in virtual museum settings or AR technologies to provide immersive audio or visual experiences based on visitor interactions.	14, 15, 17
Interactive Displays	Data collected through interactive displays or tables to track visitor interactions and provide personalized content.	16, 25
Mobile Applications	Data collected through mobile apps to track visitor movements, preferences, and interactions with exhibits.	24, 31
Audio/Video Recordings	Data collected through audio and video recordings of visitor interactions and behavior during museum visits.	14, 29

.

In their study, Abdelrahman et al. [28] examine how brain–computer interfaces (BCIs) can be utilized to create personalized experiences for museum visitors. The system utilizes EEG signals from the Emotiv EPOC device to measure real-time visitor engagement levels, which then trigger dynamic exhibit recommendations at museums. The application of BCIs shows how content can be modified according to visitors’ mental and emotional states to provide new engagement methods through neurofeedback.

Researchers frequently use questionnaires and surveys to obtain clear information about visitor preferences and demographic data. Almeshari et al. [29] collected visitor data through in-person questionnaires to develop museum visitor personas for personalized mobile guide content delivery. Rodriguez-Boerwinkle and Silvia [55] used online questionnaires to study the effect of personality traits on visitor behavior within virtual museums. Direct visitor feedback plays a crucial role in designing personalized experiences, especially for static or pre-visit customization.

Museums commonly employ computer vision techniques to monitor and study visitor behavior patterns. Castellano et al. [32] revealed that the Pepper robot uses computer vision to identify museum visitors and provide demographic information, which enables it to recommend art pieces in real-time. Ferrato et al. [34] used deep learning algorithms combined with RGB cameras to monitor visitor movements and interactions but concentrated primarily on gathering data instead of delivering real-time personalization. This research provides evidence that computer vision can be effectively used for both real-time adjustments and retrospective evaluations.

Many organizations use Bluetooth Low Energy (BLE) technology to monitor visitor movements and deliver location-based personalized experiences. Ivanov [38] combined BLE beacons with GPS to send push notifications when visitors neared exhibits and created content dynamically according to their location. Orenes-Vera et al. [48] combined BLE beacons with multimedia content to create personalized visitor experiences as they occurred.

RFID and NFC technologies allow users to engage in interactive, personalized experiences through physical contact. Hashemi and Kamps [37] implemented RFID tags to display customized content on iPad screens at specific locations. Petrelli et al. [50] utilize NFC-enabled votive lamps to produce personalized postcards that reflect visitor selections. Research demonstrates how RFID/NFC can produce tangible, personalized experiences that engage users effectively.

The use of proximity sensors enables tracking of visitor movements together with crowd behavior analysis. Martella et al. [45] implemented Zigbee-compliant proximity sensors to record visitor density and movement patterns, which helped museum management to gain insights. These studies are crowd-analysis centered, yet they show how effective proximity sensors can be at capturing visitor behavior patterns.

Multimodal sensors analyze visitor engagement by merging various data streams. Emerson et al. [33] utilized eye gaze tracking along with facial expression analysis and posture sensors to determine visitor dwell time through an all-encompassing method of engagement analysis. The research demonstrates how analyzing multimodal datasets can improve visitor experiences by allowing for in-depth behavioral examination.

Location-based personalization becomes possible in large outdoor museum settings through the application of GPS combined with geofencing technology. The research by Ivanov and Velkova [40] demonstrates how geofencing technology can send push notifications according to the visitor’s location information. Vrettakis et al. [58] implemented GPS data into the Pros-Eleusis app for generating dynamic recommendations. Research demonstrates how location-based technologies succeed in providing content that matches the specific context of users. Rashed et al. [53] implemented LIDAR sensors for examining visitor movement patterns and preferences within museum environments.

Museums use virtual and augmented reality technologies to develop immersive visitor experiences and monitor visitor behavior. Javdani Rikhtehgar et al. [41] conducted research into audience viewing habits and content preferences through VR technology. Kaghat et al. [42] created the SARIM system, which combines AR with gesture recognition to provide customized audio content. Research shows that VR and AR technologies are becoming vital tools for improving and tailoring cultural heritage experiences.

Interactive Displays serve as platforms to present customized content while collecting data on user interactions. Karaman et al. [43] developed an interactive tabletop interface designed to deliver real-time recommendations. Rajaonarivo et al. [52] used a dynamic 3D virtual museum environment to adapt content appropriately. Research findings demonstrate that Interactive Displays successfully produce personalized and engaging visitor experiences.

The delivery of personalized content through mobile apps is growing alongside their ability to monitor visitor behavior. Philippopoulos et al. [51] utilized a mobile app platform with BLE and RFID capabilities to offer real-time personalized information. In collaboration with this research, the Pros-Eleusis app used GPS data and user feedback to generate dynamic recommendations, as demonstrated by Vrettakis et al. [58]. Research findings show how mobile applications can create personalized experiences for museum visitors through their versatile capabilities.

Visitor behavior and interactions undergo analysis through the study of audio and video recordings. Roussou and Katifori [56] analyzed visitor interactions with mobile narratives through both video and audio recordings. Javdani Rikhtehgar et al. [41] gathered eye gaze and audio recordings from VR exhibitions.

3.4.2. Core Behavioral Parameters

Analyzing visitor behavior in smart museums requires determining behavioral parameters that can measure the complex experience visitors have. Our review of the existing literature led to the discovery of twelve essential behavioral parameters, which are documented in

Table 11. Core behavioral parameters.

Core Behavior	List of Sub-Behaviors
Attention Span	Focus duration, cognitive engagement, distraction patterns
Content Preferences	Media type preferences, information-seeking behavior
Dwell Time	Time spent at exhibits, visit duration, stopping patterns
Emotional Response	Affective states, engagement levels, satisfaction
Exhibit Interest	Engagement level, preference indicators, attraction power
Group Dynamics	Social clustering, group formation, collective behavior
Interactive Engagement	Hands-on interaction, digital participation, active involvement
Learning Patterns	Knowledge acquisition, comprehension, learning styles
Movement Patterns	Navigation routes, spatial distribution, flow patterns
Physical Positioning	Location tracking, proximity to exhibits, spatial behavior
Social Interaction	Social engagement, collaborative behavior
Visitor Flow	Traffic patterns, congestion points, circulation paths

.

Spatial-temporal aspects such as Movement Patterns, Dwell Time, Visitor Flow, and Physical Positioning together with cognitive-affective components like Attention Span, Emotional Response, and Learning Patterns make up the parameters. Analyzing these multiple parameters provides a detailed understanding of visitor engagement that surpasses traditional metrics such as visit duration and exhibit popularity.

Interactive Engagement, along with Exhibit Interest and Content Preferences, serves as a detailed indicator for examining qualitative visitor–exhibit relationships. Exhibit designers can develop more engaging interfaces by applying these metrics, which help distinguish passive observation from active participation. The study of Group Dynamics and Social Interaction as socially oriented parameters illuminates how collective elements and interpersonal influences help shape individual patterns of engagement within museum experiences.

Technological advancements led to significant changes in behavioral parameter quantification methods, shifting from traditional observation techniques to sensor-based automated tracking systems. The progression of research methods now allows scientists to obtain behavioral details that were once out of reach. The temporal synchronization of various behavioral parameters enables researchers to establish causal links between environmental stimuli and visitor responses, which supports personalized approaches.

3.4.3. Analysis of the Core Behaviors

Table 12. Behavioral analysis.

Behavior	Number of Papers (%)	Doc IDs
Attention Span	10 (30.3)	[1, 6, 14, 16, 17, 18, 21, 25, 28, 29]
Content Preferences	16 (48.5)	[1, 2, 3, 4, 11, 12, 13, 14, 15, 17, 20, 24, 27, 29, 30, 31]
Dwell Time	14 (42.4)	[4, 7, 9, 10, 11, 12, 13, 16, 18, 19, 21, 28, 29, 32]
Emotional Response	6 (18.2)	[1, 3, 6, 8, 14, 15]
Exhibit Interest	16 (48.5)	[4, 5, 7, 9, 10, 11, 12, 13, 15, 17, 18, 19, 21, 22, 29, 32]
Group Dynamics	9 (27.3)	[4, 7, 8, 12, 18, 19, 20, 22, 27]
Interactive Engagement	24 (72.7)	[1, 2, 5, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 19, 21, 23, 24, 25, 26, 27, 29, 30, 31]
Learning Patterns	21 (63.6)	[1, 2, 3, 5, 6, 8, 10, 11, 12, 14, 15, 16, 17, 20, 22, 23, 24, 25, 26, 28, 30]
Movement Patterns	24 (72.7)	[2, 4, 5, 6, 7, 9, 10, 11, 12, 13, 15, 16, 18, 19, 20, 21, 22, 24, 25, 26, 28, 29, 31, 32]
Physical Positioning	21 (63.6)	[5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 18, 19, 20, 21, 22, 23, 24, 26, 28, 31, 32]
Social Interaction	13 (39.4)	[5, 7, 8, 16, 18, 19, 22, 24, 26, 27, 28, 29, 33]
Visitor Flow	9 (27.3)	[4, 9, 18, 19, 20, 22, 24, 26, 32]

provides a detailed breakdown of the number of publications that focused on each of the 12 core behavioral parameters, along with the corresponding document IDs.

Most Frequently Used Behaviors

The behaviors of Interactive Engagement and Movement Patterns appeared as the most frequently studied topics, with each having been covered in 24 publications, which constituted 72.7% of the total research. The data show a clear focus on studying visitor interactions with exhibits and their movement patterns within museum spaces because these elements are essential for creating individualized and captivating visitor experiences. A total of 21 publications focused on Learning Patterns and Physical Positioning, indicating that both topics received extensive research attention, with 63.6% of publications each. Visitor preferences and spatial behavior analysis help in optimizing exhibit placement and content delivery.

Moderately Analyzed Behaviors

The research on Content Preferences and Exhibit Interest appeared in 16 publications, representing 48.5% of the total for each topic. Analyzing visitor behaviors enables museums to understand which content and exhibits draw visitors and helps them design more engaging and relevant experiences. Fourteen publications (42.4%) investigated Dwell Time because it measures visitor duration at exhibits to improve exhibit design and visitor flow management.

Less Frequently Analyzed Behaviors

Research conducted across 10 publications (30.3%) analyzed Attention Span to understand visitor engagement duration with exhibits and content. The research showed that 27.3% of the studies focused on Group Dynamics and Visitor Flow, which implies these behaviors play significant roles in social interaction and traffic management, but they do not often serve as central research topics. A mere six publications (18.2%) investigated Emotional Response, marking it as the behavior with the fewest research studies. The data reveal that despite the recognized importance of affective states and engagement levels in museum visits, they are seldom measured or used to create personalized experiences.

3.4.4. Meta-Analysis

A meta-analysis of 33 publications on visitor behavior in museums reveals significant disparities in research focus, with Interactive Engagement and Movement Patterns being the most studied (24 publications each, 72.7%), while Emotional Response received the least attention (6 publications, 18.2%). The Chi-square test (${\chi }^2$ = 27.4, df = 11, p < 0.01) demonstrates that the distribution of research across behaviors is highly uneven, rejecting the null hypothesis of equal coverage. These findings highlight a strong research bias toward quantifiable, interaction-based behaviors like exhibit engagement and visitor movement, with comparatively less emphasis on emotional and social aspects, suggesting a need for future studies to address this imbalance for a more comprehensive understanding of museum visitor experiences.

3.5. Personalization Technologies and Strategies

The review of 33 publications showed that 19 papers focus on ways museums deliver personalized experiences to their guests. The publications chosen for analysis fall into 14 categories according to the delivery methods and technologies for personalized content to museum visitors. The groups are named and described in

Table 13. Personalization group names.

Group Name	Description	Paper IDs
Real-Time Personalization	Content is delivered in real-time based on visitor behavior or preferences.	1, 2, 5, 8, 10, 11, 12, 13, 21, 24, 25, 27, 29, 30, 31
Pre-Visit Personalization	Content is personalized based on pre-visit data or initial visitor input.	2, 3, 10, 15, 29
Post-Visit Personalization	Content is personalized after the visit based on collected data.	16, 23
Static Personalization	Content is personalized based on static categories or predefined personas.	3, 15
Dynamic Personalization	Content is dynamically updated during the visit based on real-time data.	1, 5, 8, 11, 12, 13, 21, 24, 25, 27, 29, 30, 31
Location-Based Personalization	Content is personalized based on the visitor’s location in the museum.	11, 13, 21, 24, 31
Emotion-Based Personalization	Content is personalized based on detected emotional states of visitors.	1, 8
Interaction-Based Personalization	Content is personalized based on visitor interactions with exhibits or interfaces.	5, 10, 16, 23, 25, 27, 29
AI-Driven Personalization	Content is personalized using AI algorithms and machine learning.	1, 8, 11, 12, 30, 31
Sensor-Based Personalization	Content is personalized using data from various sensors (e.g., BLE, RFID/NFC).	1, 5, 10, 11, 12, 13, 21, 24, 27
Questionnaire-Based Personalization	Content is personalized based on data collected from questionnaires.	2, 3, 10, 29
Virtual/Augmented Reality Personalization	Content is personalized using VR/AR technologies.	14, 15, 17, 25, 28
Chatbot-Based Personalization	Content is personalized through interactions with a chatbot.	11, 30
No Personalization	No personalized content is delivered to visitors.	4, 6, 7, 9, 14, 17, 18, 19, 20, 22, 26, 28, 32, 33

, which also lists publication identifiers for each group.

Abdelrahman et al. [28] studied a brain–computer interface system that monitors engagement through EEG signals to deliver personalized museum exhibit suggestions during real-time visits. Similarly, Castellano et al. [32] introduced the Pepper robot, which utilizes computer vision technology to recognize museum visitors and provide real-time art recommendations as they interact with exhibits. Ivanov’s system, ExhibitXplorer [38], integrates geofencing with artificial intelligence and push notifications to dynamically deliver customized content to museum visitors as they move through the exhibit space. Research demonstrates that adaptive technologies for real-time personalization create improved visitor engagement opportunities.

Personalizing content for museum visits can start by collecting visitor data in advance through surveys or initial input forms. In their study, Almeshari et al. [29] collected visitor information using questionnaires that captured motivations and success criteria, enabling the system to generate personas and deliver customized experiences. Hashemi and Kamps [37] utilized a check-in station where visitors could record their preferences to customize the content presented at each Point of Interest (POI) during their visit. The research findings validate how early visitor input can generate customized museum experiences through pre-visit personalization techniques.

Another personalization technique involves tailoring content after the visit by analyzing collected data. During their visit, visitors interact with a tabletop interface to access personalized content, which the study by Karaman et al. [43] presents, while visitors can obtain extra personalized content through a QR code on their mobile devices after their museum tour. Petrelli et al. [50] presented an interactive installation where visitors used NFC-enabled votive lamps to engage with stands, resulting in the creation of a personalized postcard upon completion of their interactions. The described methods demonstrate that museums can extend visitor engagement through personalized experiences that continue after the visit.

Multiple studies examine static personalization, which adjusts content through predefined categories or personas. Almeshari et al. [30] used initial question responses to define personas and preferences, which then personalized content delivery. The research by Kaghat et al. [42] applied static personalization to modify content according to visitor categories, either before or after their visit. These methods display how fixed categories enable personalized experiences without requiring real-time changes.

Research shows that dynamic personalization, which updates content during user visits using real-time data, stands as a central topic of investigation. The META4RS system, introduced by Ferrato et al. [35], allows for the creation of tailored visitor itineraries that change based on both detected visitor interests and their emotional responses to art pieces. The authors Rajaonarivo et al. [52] presented a real-time evolving 3D “living” museum metaphor that recommended objects according to user profiles through interactive engagement. Research demonstrates that dynamic personalization holds promise for developing museum experiences that are both adaptive and engaging.

Personalization at museums is achieved by adapting content to match visitors’ real-time locations. The study by Ivanov [38] utilizes geofencing alongside GPS technology to send personalized push notifications to visitors as they approach specific museum exhibits. BLE beacons enable the monitoring of visitor movement patterns and deliver location-based multimedia content, according to Orenes-Vera et al. [48]. Studies demonstrate that location-aware technologies enhance personalization by delivering content specific to users’ contexts.

Studies have evaluated emotion-based content delivery systems that adapt to users’ emotional states through a few research investigations. Abdelrahman et al. [28] examined EEG signals to determine engagement levels, enabling them to provide tailored recommendations. Ferrato et al. [35] analyzed facial expressions to deliver adaptive recommendations that respond to users’ emotional states in real time. Emotion-aware systems show their capability to improve museum experiences by providing empathetic and engaging interactions through these innovative approaches.

Personalization that adapts content through user interaction with exhibits and interfaces stands as a crucial focus area. Castellano et al. [32] used computer vision technology to monitor visitor behavior with artworks and modify the exhibited content based on this data. Rey et al. [54] introduced a tangible interface which modifies visit routes in real-time according to visitor interactions. The research demonstrates how interactive technologies facilitate the creation of both personalized and engaging experiences for museum visitors.

The field of research which utilizes AI-driven personalization through machine learning and artificial intelligence is expanding rapidly. The research by Ivanov [39] presents BLE smart badges which use telemetry data to enable real-time visitor tracking and group dynamics that support AI-based personalized content delivery. Tsitseklis et al. [57] developed a chatbot which combines Natural Language Processing (NLP) capabilities with a hybrid recommender system to deliver real-time exhibit suggestions. Research shows that AI systems can develop sophisticated personalization techniques that adapt intelligently to museum environments.

A significant number of studies (14) focus on data collection and analysis without delivering personalized content. These studies emphasize the importance of understanding visitor behavior for museum management and exhibition optimization, even in the absence of personalization.

4. Discussion

4.1. Research Question 1

This section addresses RQ1: What are the current methodologies for analyzing visitor behavior in smart museums?

4.1.1. Findings

The review of methodologies used to examine visitor behavior in smart museums demonstrates an evident shift towards combining various methods while focusing heavily on quantitative techniques. Statistical and Data Analysis (SDA) became the leading methodology and appeared in 97% of the reviewed papers, which demonstrates its essential role in visitor behavior analysis (Table 6). The widespread adoption of data analysis methods in this field reflects a strong dedication to utilizing data-driven methods for creating tailored museum experiences. The high occurrence of SDA combined with AI and Machine Learning (AML) in 63.6% of research papers demonstrates an increased dependence on sophisticated predictive analytics to decode complex visitor behavior patterns (Table 7). Research from Ferrato et al. [34] and Ivanov [38] shows how artificial intelligence improves personalized content delivery accuracy.

The studies revealed that Mobile and Interactive Technologies (MIT) appeared in 60.6% of them while frequently being implemented alongside SDA. Almeshari et al. [30] and Kaghat et al. [42] demonstrated how real-time interaction mapping helped researchers analyze visitor engagement with museum exhibits. Both Geospatial Methods (GSM) and Survey and Interview-Based Methods (SIB) received moderate application, according to research by Ivanov and Velkova [40] and Petrelli et al. [50]. These approaches deliver comprehensive insights into visitor behavior by tracking both their physical actions and personal reactions.

A multi-methodology research design trend emerged from the review, as 39.4% of papers utilized four separate methodologies. Research following this approach includes the work of Rashed et al. [53]. The research field acknowledges visitor behavior’s complex nature, which necessitates the integration of technological, spatial, and psychological perspectives. The limited use of complex research designs involving 6–7 methodologies indicates practical implementation challenges identified in the methodological complexity analysis. The multi-methodology strategy improves predictive analytics precision while helping create engaging, personalized museum experiences for visitors. Emerging technologies, including Virtual/Augmented Reality and Biometric Monitoring, require further research to improve existing methodologies and solve current approach limitations.

4.1.2. Limitations and Future Trends

Statistical and Data Analysis (SDA) and AI and Machine Learning (AML) focus predominantly on quantitative information yet fail to capture detailed qualitative visitor aspects like emotional engagement and cultural context, which Survey and Interview-Based Methods (SIB), alongside Biometric and Physiological Monitoring (BPM), can better document. The combination of diverse methodologies provides advantages yet demands substantial resources and expertise, which smaller institutions may find inaccessible. The use of Wireless Signal-Based Tracking (WST) and Computer Vision and Sensor-Based Methods (CVS) presents privacy issues because they gather sensitive data about visitors. The next developments in this field will likely concentrate on bridging current limitations through ethical data practices and scalable multi-method solutions alongside the integration of Virtual/Augmented Reality (VAR) and sophisticated biometric systems. These technological advancements will enable smart museums to deliver personalized and immersive visitor experiences that preserve personal privacy and foster comprehensive insights into visitor behavior.

4.2. Research Question 2

This section addresses RQ2: Which technologies and the analysis of what museum visitor behaviors contribute to content personalization?

4.2.1. Technologies

Findings

Technologies with Behavioral Analytics Platforms received the highest number of citations at 14, which represents 42.4% of all research papers. Through visitor analytics and heatmaps, along with path analysis, Behavioral Analytics Platforms become essential tools for understanding visitor behavior. For instance, Martella et al. [47] applied behavioral analytics to chart visitor behavior patterns and make predictions about museum visitor interactions which showed trends in exhibit engagement. Similarly, Ferrato et al. [34] gathered visitor behavior data at museums using deep learning techniques and demonstrated the application of AI-driven analytics to understand visitor engagement patterns.

The research on Mobile and Wearable Devices showed equal prominence, with appearances in 14 papers, which accounts for 42.4%. Museums utilize smartphones, tablets, and wearable sensors to provide personalized experiences and visitor guidance throughout exhibits. Ivanov [38] created Exhibitxplorer, which utilizes mobile devices and geofencing to provide museum visitors with personalized content. Additionally, Kaghat et al. [42] investigated mobile device interfaces that utilize gesture-based sound augmented reality to enrich museum experiences.

According to 13 papers, which make up 39.4% of the research, Location Tracking Systems serve as the primary technology to monitor visitor movements and provide context-aware content. Handojo et al. [36] applied an indoor positioning system for monitoring museum visitor activities, which allowed location-based content delivery. The research by Ivanov and Velkova [40] expanded the use of geofencing to provide personalized content specifically for visitors at open-air museums.

AI/ML Systems feature in 11 papers, which is 33.3% of the total, showing their critical role in developing personalized museum experiences with recommender systems and personalization engines. The META4RS system was proposed by Ferrato et al. [35] to utilize AI for analyzing visitors’ emotions and behaviors to create personalized recommendations. The study by Hashemi and Kamps [37] applied AI technology to create behavioral user models which help in making point-of-interest recommendations in smart museums.

A total of 11 papers (33.3%) utilize Survey/Feedback Systems to demonstrate their function in gathering visitor feedback for enhancing museum experiences. Almeshari et al. [30] performed research on preferences for museum mobile guides by analyzing survey data to determine visitor persona requirements. Rey et al.’s [54] development introduced a museum experience system that permits visitor personalization through feedback-based content customization.

Nine papers (27.3%) feature Computer Vision Systems because they track and analyze visitor behavior using RGB cameras, facial detection, and person-tracking technologies. The SeSAME system, which utilizes computer vision technology for re-identification and ambient intelligence functions in museum contexts, was developed by Paolanti et al. [49]. The research by Rashed et al. [53] employed computer vision technology to monitor visitor behavior inside a real museum.

Content Management Systems feature in nine papers (27.3%), where they enable museums to provide visitors with personalized content delivery. The system developed by Karaman et al. [43] uses the passive observation of museum visitors to deliver personalized multimedia content through an interactive table. Rajaonarivo et al. [52] introduced an evolving museum metaphor to enable personalized content delivery through Content Management Systems that adjust to visitor interests.

Nine papers (27.3%) utilized IoT Sensors to gather visitor proximity and environmental condition data that improve the museum visitor experience. Through proximity sensing, Martella et al. [46] examined museum visitor behavior, which showed how IoT technology can analyze visitor interactions. Petrelli et al. [50] examined IoT applications at heritage sites and demonstrated how they can enable interactive and personalized visitor experiences.

Five papers (15.2%) implemented Biometric Sensors to analyze visitor engagement using eye tracking and heart rate monitoring technologies. Through brain–computer interfaces, Abdelrahman et al. [28] identified implicit visitor engagement in museums, which demonstrates biometric data’s potential to analyze visitor behavior. Javdani Rikhtehgar et al. [41] incorporated Biometric Sensors in virtual reality exhibitions to tailor cultural heritage experiences according to visitors’ viewing patterns.

The implementation of Virtual/Augmented Reality features in five scholarly papers (15.2%) demonstrates the growing focus on immersive museum experiences through these technologies. Liu and Sutunyarak [44] examined the effects of immersive technology on visitor behavioral intentions, demonstrating how VR and AR can enhance engagement. The study by Rodriguez-Boerwinkle and Silvia [55] demonstrated how visitors’ behavior in virtual museums is affected by personality traits while offering valuable information for VR experience customization.

Five papers (15.2%) describe the use of Interactive Displays to create engaging visitor interactions with museum content. Karaman et al. [43] and others successfully demonstrated Interactive Displays’ potential to enhance visitor engagement using interactive tables for personalized multimedia delivery. Roussou and Katifori [56] conducted a user experience study on mobile museum narratives, which included Interactive Displays to enable personalized storytelling.

Social Media Integration appears in just two papers and represents only 6.0% of the reviewed technology usage. For example, Zheng et al. [60] conducted research to identify elements that drive user ongoing engagement in digital museums while including social media analysis in their study. Social media shows potential for engaging museum visitors, but its minimal presence in reviewed studies reveals it remains an uncommon tool for personalization in smart museums.

Limitations and Future Trends

Behavioral Analytics Platforms, alongside other tracking technologies, face the major challenge of ensuring visitor data privacy and security. When Behavioral Analytics Platforms collect sensitive information, such as visitor movement patterns and biometric details, along with personal preferences, they create ethical challenges. For example, Abdelrahman et al. [28] developed brain–computer interfaces which gathered sensitive personal data to measure user engagement. Future researchers must develop privacy protection solutions which sustain personalized experiences by investigating anonymization technologies and establishing transparent data usage policies.

Integrating various technologies together, like AI/ML Systems with IoT Sensors or Virtual/Augmented Reality, continues to pose significant challenges. The META4RS system introduced by Ferrato et al. [35] combines behavioral analytics with AI and Biometric Sensors, even though this kind of integration remains uncommon. Future investigations need to identify methods for integrating these technologies to build comprehensive and unified visitor experiences.

The implementation of Interactive Displays along with Virtual/Augmented Reality and Biometric Sensors demands substantial hardware and software investments. Liu and Sutunyarak [44] identified immersive technology as a promising development yet acknowledged that the financial demands for VR headsets and AR applications could restrict their adoption in smaller museum settings. Researchers need to develop affordable solutions that can expand across various museum sizes, including both major institutions and smaller venues with limited resources.

Personalized museum experiences become successful when visitors accept them and find them easy to use. Mobile and Wearable Devices along with Interactive Displays need to operate intuitively and remain user-friendly for effective use. Researchers should prioritize developing technologies that visitors of all technical backgrounds can easily access.

The discussion of both challenges and future possibilities illustrates why persistent research and inventive progress are essential for smart museum domains. Behavioral Analytics Platforms alongside AI/ML Systems and Location Tracking Systems have improved personalized visitor experiences even with ongoing challenges related to data privacy, scalability, and visitor acceptance. Research teams need to merge diverse technologies as they develop new VR/AR applications which adhere to ethical standards and inclusive guidelines.

4.2.2. Behaviors

Findings

Twelve essential behavioral parameters emerged from the systematic review for visitor behavior analysis and modeling in smart museums.

Visitor interaction with exhibits through hands-on activities, digital participation, and active involvement was the primary focus of 24 publications (72.7%) analyzing Interactive Engagement as the most studied visitor behavior. Visitor interaction is essential to develop captivating and immersive experiences in museums. The primary publications that advanced this research area comprise work by Abdelrahman et al. [28], who researched brain–computer interfaces within museum environments to study visitor engagement and offered technological insights to improve interactive experiences. The work of Emerson et al. [33] explored early visitor engagement prediction through multimodal learning analytics, which resulted in establishing a framework for immediate engagement evaluation. Additionally, Kaghat et al. [42] introduced a gesture-based sound augmented reality interface for museums, which showed how interactive technologies enhance visitor participation.

The analysis of Movement Patterns included 24 publications (72.7%) which focused on visitor navigation within museum spaces. Knowing visitor movement patterns helps museum planners optimize exhibit locations and control visitor traffic. Martella et al. [45] demonstrated through density measurements how human behavior patterns emerge in crowded spaces while providing valuable information about visitor movements in busy areas. Yoshimura et al. [59] used random walk models to study visitor movement patterns within the Louvre Museum, which provides a new methodological framework for analyzing spatial behavior. Furthermore, Orenes-Vera et al. [48] developed the RECITE framework to analyze user trajectories at cultural sites and used it to deliver personalized content in museums.

The analysis of Learning Patterns across 21 publications (63.6%) underscored the significance of understanding visitor knowledge acquisition and educational content engagement. To design museum exhibits that cater to various learning styles, it is essential to understand visitor behavior. Almeshari et al. [30] investigated mobile guide preferences among various visitor personas, providing insights into enhancing learning experiences through personalized content. The study by Karaman et al. [43] analyzed how personalized multimedia delivery through passive visitor observation enables adaptive systems to support learning. Rashed et al. [53] conducted a study which involved tracking actual museum visitors to examine behavioral patterns linked to learning and engagement.

The field of Physical Positioning received attention through 21 publications (63.6%), which highlighted the critical need to monitor visitor locations and their distances from exhibits. Such behavior enables museums to provide content that responds to the visitor’s location while optimizing how they present their physical spaces. Paolanti et al. [49] developed the SeSAME system, which utilizes re-identification techniques to support ambient intelligence in museums by enabling precise visitor location tracking. Handojo et al. [36] implemented a museum visitor activity tracking system through indoor positioning technologies to showcase practical location-based applications. Ivanov [39] introduced advanced visitor profiling methods with telemetry-driven smart badges that create personalized museum experiences through physical positioning-based tracking.

Limitations and Future Trends

Emotional Response received the least amount of research attention, with just six studies, representing 18.2%, investigating affective states and engagement levels. Affective computing technologies, such as facial expression analysis, heart rate monitoring, and sentiment analysis, should be incorporated into future studies to improve our understanding of visitor emotions and our ability to respond to them. For example, Abdelrahman et al. [28] showed brain–computer interfaces could effectively detect engagement yet require further research to apply these results to emotional responses. Visitor satisfaction and engagement will benefit greatly from systems designed to modify content according to real-time emotional feedback.

Nine publications (27.3%) addressed Group Dynamics, while thirteen publications (39.4%) analyzed Social Interaction. The behaviors visitors exhibit toward one another serve as fundamental elements for comprehending both their interactive patterns and the social influences on their museum experiences. Researchers should explore how collaborative technologies can improve group dynamics through shared augmented reality (AR) and virtual reality (VR) environments. For instance, Rashed et al. [53] conducted research on group visitor behavior, yet more investigation is required to determine technological solutions that enable social learning and interaction within museum environments.

A range of studies investigated immersive technology applications in augmented reality (AR), virtual reality (VR), and mixed reality (MR), with research by Kaghat et al. [42] and Liu and Sutunyarak [44]. There are numerous research opportunities in creating strategies that incorporate these technologies to build personalized and interactive experiences for museum visitors. Future research needs to focus on the long-term impact of immersive technologies on visitor learning and engagement while addressing ethical issues surrounding data collection in these environments.

Although visitor behavior analysis has been a focus of many studies for personalization research, there is potential for further development with artificial intelligence (AI) and machine learning (ML). The use of natural language processing (NLP) to create personalized recommendations, which Tsitseklis et al. [50] demonstrated, opens new possibilities for AI-driven content personalization. The concept of natural language processing systems for personalized recommendations should be extended to create complex conversational agents and chatbots.

Most of the examined studies examined brief interactions or took place within regulated settings. Longitudinal studies that monitor visitor behavior over timeframes would help future research understand the evolution of personalized experiences. To validate proposed technologies and methodologies, future research must include more real-world applications and case studies.

Cross-disciplinary collaboration drives the advancement of smart museums through combined contributions from computer science experts, psychologists, educators, and cultural heritage specialists. Rodriguez-Boerwinkle and Silvia [55] researched the impact of personality traits on virtual museum visits, which emphasized the possibility of using psychological theories to enhance museum design. Researchers should investigate how cognitive science combined with human–computer interaction and cultural studies can guide the creation of better and more inclusive museum experiences.

4.2.3. Technology–Behavior Correlation Patterns

Analysis of the relationship between twelve distinct technological solutions and twelve visitor behavior metrics reveals several significant patterns (see Figure 9).

The analysis reveals four separate technology clusters that are grouped according to their main behavioral domains:

• Spatial Tracking Technologies: Spatial behavioral metrics are dominated by Location Tracking Systems together with Computer Vision Systems, which build a strong technological base for analyzing visitor movement patterns as well as positioning and flow.
• Content Engagement Technologies: Content Management Systems paired with Survey/Feedback Systems and AI/ML Systems focus on cognitive engagement while enabling personalized content delivery and learning assessment.
• Interactive Experience Technologies: Mobile and wearable devices exhibit a strong link to interactive engagement (n = 9), which makes them crucial for enabling visitor participation.
• Behavioral Analytics Technologies: Behavioral Analytics Platforms maintain balanced metrics distribution while functioning as integrative systems that combine data from various behavioral domains.

The relationship between AI/ML Systems and Learning Patterns (n = 8) and Content Preferences (n = 6) demonstrates their ability to effectively adapt to visitor learning behaviors and content interests. AI/ML Systems demonstrate significant potential in customizing educational materials to match the preferences of individual visitors. Despite their strengths, AI/ML Systems demonstrate weak connections to Emotional Response and Social Interaction and lack ties to Visitor Flow, which points to limitations in emotional and social dynamics understanding that could necessitate integration with additional technologies.

Visitor movement and exhibit-engagement tracking capabilities are demonstrated by Behavioral Analytics Platforms through their moderate connections to Dwell Time (n = 4), Exhibit Interest (n = 4), and Physical Positioning (n = 5). These systems demonstrate limited connection with Interactive Engagement (n = 1) and Learning Patterns (n = 2), indicating their effectiveness lies in physical behavior monitoring but falls short in capturing interactive or cognitive visitor engagement elements.

Through demonstrated associations with Attention Span (n = 4) and Emotional Response (n = 4), Biometric Sensors show promise in capturing instantaneous physiological and emotional data. The restricted linkages between Biometric Sensors and Group Dynamics and Social Interaction metrics suggest they function better for single-person evaluations instead of group assessments.

Movement Patterns demonstrate strong connections with Computer Vision Systems (n = 8), and these systems maintain moderate relationships with Social Interaction (n = 5) and Physical Positioning (n = 6). Through the tracking of visitor movement and interaction within museums, these systems demonstrate their potential to analyze spatial behavior patterns as well as social dynamics. The people counting system developed by Holla et al. [63] attained 94.2% accuracy during experiments with the TUD-Pedestrian dataset and demonstrated a perfect accuracy of 100% when applied to real-world settings. Their lower correlation with Emotional Response and Learning Patterns indicates these systems may fall short in completely detecting cognitive or emotional states.

Location Tracking Systems achieved leading positions in Movement Patterns (n = 11) and Physical Positioning (n = 12) benchmarks, validating their excellent performance for visitor movement tracking and spatial positioning. The practical capabilities of the system enable successful optimization of exhibit arrangements along with management of visitor traffic flow. Location Tracking Systems lack the ability to integrate Emotional Response and Content Preferences metrics, which shows that new technologies are required to manage these specific components.

Interactive Engagement matches well with Mobile and Wearable Devices (9), but these devices show moderate connections to Content Preferences (1), Movement Patterns (4), and Social Interaction (4). Current analysis reveals Mobile and Wearable Devices can enable interactive visitor experiences alongside preference data collection in their unique formats. The limited relationship between these technologies and Emotional Response, as well as Group Dynamics, indicates room for future improvement.

The Survey/Feedback Systems show the highest relationship with Content Preferences (n = 10), which validates their ability to gather clear visitor feedback on content. The moderate relationships between these factors with Learning Patterns (n = 5) and Dwell Time (n = 1) show they can provide valuable information about visitor engagement and learning outcomes. The systems fail to link Movement Patterns and Social Interaction, which shows their limitations in acquiring real-time or behavioral data.

Technologies such as IoT Sensors and Virtual/Augmented Reality establish moderate connections with performance metrics but fail to provide comprehensive coverage for all behavior types. IoT Sensors display moderate associations with Physical Positioning (n = 4) and Social Interaction (n = 3), whereas Virtual/Augmented Reality demonstrates limited yet significant links to Interactive Engagement (n = 4) and Movement Patterns (n = 2). These systems function as adjunct tools instead of complete solutions in delivering personalized content.

Visitor behavior analysis shows that existing technological offerings fall short of addressing every analytical requirement. An all-encompassing personalized visitor experience requires technological integration, including AI/ML Systems for content preference analysis, Biometric Sensors for emotional response tracking, and Location Tracking Systems to assess movement patterns. Further research should explore integrated technological systems that function through combined strengths to address their individual limitations. Ethical considerations, including privacy protection and data security, should stay at the forefront of smart museum environments that implement these technologies.

4.3. Research Question 3

This section addresses RQ3: Which personalization technologies and strategies are most used and promising?

4.3.1. Findings

Modern museums are using artificial intelligence (AI), computer vision, and sensor-based systems to create personalized experiences for their visitors. The study identifies key findings, which we explore by analyzing their effects on museum operations and visitor experiences together with outlining future research directions.

Brain–computer interfaces (BCIs) and computer vision combined with geofencing technologies represent the principal approach in delivering adaptive content during real-time visitor experiences. For instance, Abdelrahman et al. [28] demonstrated the capability of BCIs to monitor visitor engagement levels and provide real-time suggestions. The work by Castellano et al. [32] explained how the Pepper robot provided customized artwork recommendations during visitor interactions. The systems enhance visitor participation while providing museum patrons with immersive and interactive experiences. The necessity for real-time data processing generates multiple challenges, which feature computational efficiency issues and privacy concerns together with the requirement for robust supporting infrastructure.

Almeshari et al. [29] and Hashemi and Kamps [37] research demonstrated how pre-visit personalization works. Initial visitor input from questionnaires and preference settings allows for content customization before their arrival. Museums can start personalized experiences immediately through this method, yet they must have accurate data collection and effective profiling systems to succeed. Post-visit personalization, exemplified by Karaman et al. [43] and Petrelli et al. [50], extended museum experiences beyond physical visits by providing personalized content through digital summaries and postcards based on visitor data collection. Through the adoption of these strategies, museums maintain visitor engagement after their physical visit, which leads to enduring audience connections.

Access to real-time data lets museums implement dynamic personalization, which leads the way in personalization trends by constantly updating visitor content during their experience. The research by Rajaonarivo et al. [52] examines the adaptive capabilities of recommendation systems which respond to visitor behavior and emotional feedback to match their interests. Ivanov [38] and Orenes-Vera et al. [48] showed how location-based personalization operated. Location-based personalization systems deliver visitor-specific information according to their current position within the museum to improve adaptability. The described approaches prove that context-aware systems play a crucial role in delivering seamless and engaging visitor experiences.

Emotion-based personalization, explored by Abdelrahman et al. [28] and Ferrato et al. [35], targeted human emotions to customize content based on continuous emotional monitoring. This method enables museums to design experiences which enhance emotional engagement and empathy between visitors and exhibits. Castellano et al. [32] provided an example of interaction-based personalization. Research shows that examining visitor behavior in real time plays an essential role in developing engaging museum spaces where visitors can participate actively.

Many researchers have focused on AI-driven personalization, which continues to expand rapidly, as shown by Ivanov [39] and Tsitseklis et al. [57]. Advanced personalization systems combine machine learning techniques with natural language processing (NLP) and telemetry inputs to deliver real-time suggestions and automatic content modification. AI integration into museum personalization creates opportunities for ultra-personalized experiences yet presents ethical challenges, including data privacy concerns and algorithmic bias, which necessitates the development of transparent AI systems.

4.3.2. Limitations

The use of personalized systems in smart museums creates important ethical questions about how data are collected and how user consent is obtained, which demand thorough evaluation to maintain a balance between technological advancements and visitor rights.

Research results show substantial progress in museum personalization while ethical challenges continue to exist. Operating advanced technological systems and processing data in real time present substantial challenges for smaller museums because they require both extensive computing power and reliable infrastructure. Personalized experiences become accessible only through advanced technology, which creates a gap between larger and smaller cultural institutions in terms of technological equality.

Ethical challenges in data collection are paramount. Facial recognition technology [34], Biometric Sensors [28] and Location Tracking Systems [39] acquire sensitive personal information, including emotional states and physiological responses, without obtaining comprehensive explicit permission. Systems usually function by presuming visitors’ consent through their passive engagement since they operate based on an opt-out framework rather than opt-in. The method described violates the fundamental standards of informed consent and transparency that should govern personal data processing.

Present systems fail to differentiate between types of data, such as anonymized analytics and identifiable records, as well as intended data usage, such as real-time personalization and long-term research. Under GDPR and equivalent laws, specific consent requirements must be observed, yet museum technologies frequently consolidate permissions into broad service terms agreements. Existing systems fail to provide sufficient transparency about data collection while notifying visitors about its nature and extent.

Real-time adaptive systems (e.g., AI-based recommendations) need ongoing consent options so visitors can adjust their permissions throughout their visit. Privacy management interfaces must become more intuitive and adaptable so they can provide both convenience and protection of personal data.

When AI personalization systems use training data that overrepresent dominant visitor demographics, they risk reinforcing existing biases. When systems need active opt-in from users, they risk pushing privacy-focused visitors away from personalized experiences and create an engagement model where only consenting users receive full benefits. The ethical challenge emerges from ensuring that all visitors receive equal access to cultural content regardless of their personal characteristics.

Research objectives within this domain need to establish scalable personalization systems suitable for multiple museum settings which operate within present-day ethical guidelines. Additional studies must connect human–computer interaction with psychology and museum studies to create visitor-focused personalization methods. AI personalization demands comprehensive ethical evaluation to secure fairness and transparency while obtaining visitors’ informed consent.

Longitudinal studies should be used to evaluate how personalization technologies affect visitor engagement and satisfaction because these studies need to examine both the immediate and long-term impacts on museum experiences and educational results. Museums need to create sufficient privacy measures to uphold ethical transparency during visitor information collection while maintaining a balance between technological advancements and privacy protection.

4.4. Implementation Challenges

The analysis of disadvantages reveals several continuous issues and limitations present in academic research concerning personalized content delivery at museum sites. The recognized issues guide future research and development efforts toward improving system effectiveness and implementation.

The generalizability of findings remains limited because of small sample sizes. For example, Abdelrahman et al. [28] used a small sample size of 10 participants, which might fail to represent actual conditions in museum settings. Similarly, Javdani Rikhtehgar et al. [41] used 31 participants, thus reducing the generalizability of their findings. Subsequent studies must focus on expanding participant diversity and sample size to achieve results that are representative and applicable across various museum contexts.

The use of self-reported data presents a major challenge because it opens the way for subjective bias. The study conducted by Almeshari et al. [29] depended on self-reported information, which might lead to potential bias. According to Rodriguez-Boerwinkle and Silvia [55], self-reported personality measures can lead to biases and degraded data quality, thus reducing the reliability of their study findings. Researchers should consider using objective data collection techniques like sensor-based tracking and multimodal data analysis to supplement self-reported data in future studies.

The implementation of personalized systems faces challenges due to environmental conditions and technical restrictions. The research by Angeloni et al. [31] showed that system performance declined because they encountered inadequate lighting conditions and lacked modern infrastructure due to the building’s historical designation. The system presented by Castellano et al. [32] showed low accuracy in age estimation, which was affected by changing lighting conditions. Handojo et al. [36] found their Bluetooth Low Energy (BLE) beacon-based system’s accuracy was significantly influenced by environmental conditions, leading to visitor tracking inaccuracies. Janczak et al. [64] noted that the system produced an average location error of 2–3 m for visitors, as influenced by receiver distribution and environmental characteristics. The positioning system records a maximum error of 5.5 m because metal structures and other interference sources exist. Chen et al. [65] performed comparable institutional building experiments and achieved localization accuracy rates as high as 86% in certain scenarios through the combined use of Wi-Fi and BLE networks. Researchers need to direct future efforts towards creating stronger technologies which maintain functionality across multiple demanding settings.

Proactive strategies must guide the ethical integration of customized museum systems to balance technological advances with institutional obligations and visitor rights. All projects must comply with existing data protection legislation and proven standards of protection. While the GDPR and the EU’s Artificial Intelligence Act provide basic principles for data protection, museums need specialized standards to suit their unique operating environments. Visitor-centered design principles establish basic methods for developing personalized museum experiences that follow ethical guidelines. In the current design of systems for delivering personalized content, privacy is seen as a secondary factor rather than being placed at the core of the design. New data collection systems must build transparency by offering detailed explanations of the data collected and their role in enhancing personalized experiences. Systems should inform visitors that they are using personal data, e.g., “We are analyzing your itinerary to offer you exhibitions”, along with options for users to stop these data being collected. Smaller institutions find it difficult to develop ethical personalization systems themselves and thus experience unequal access to technology. Museums can circumvent these limitations by developing strategic partnerships that leverage anonymous shared datasets for AI training. Institutions with limited budgets may benefit from open-source toolsets that help manage consent procedures, detect biases in systems, and provide secure data anonymization. The EU’s Cultural Heritage Cloud project is creating a funding framework to launch ethical technology pilot initiatives in museums of different sizes. Through its workshops on the application of AI and shared resources, the Museums + AI Network demonstrates the practical application of collaborative approaches.

Numerous studies fail to conduct thorough assessments of genuine user engagement and satisfaction levels. Ivanov and Velkova [40] recognized that their research paper did not perform an extensive assessment of actual user engagement and satisfaction because it used simulation environments instead. Philippopoulos et al. [51] stated that the system’s ability to improve visitor experience and museum performance is undergoing evaluation. The pilot study conducted by Rey et al. [54] had a small sample size and short duration that may limit its ability to assess long-term system effectiveness. Longitudinal evaluations combined with real-world testing should be used in future studies to examine how personalized systems maintain visitor engagement and satisfaction over time.

The analyzed publications fail to address the financial requirements of implementing the proposed systems, which represents a major deficiency. The lack of budgetary information makes these systems impractical and unfeasible for museums that operate with restricted financial resources. Several suggested systems depend on technologies including brain–computer interfaces (BCIs), Bluetooth Low Energy (BLE) beacons, and AI-powered recommendation systems, as well as computer vision. These innovative personalization technologies for visitor engagement carry significant initial and ongoing maintenance expenses. For example, Abdelrahman et al. [28] introduced a system that utilized EEG-based BCIs and depended on specialized hardware such as the Emotiv EPOC headset. The system presented by Ferrato et al. [34] combines deep learning with computer vision but demands substantial computational resources and infrastructure support. The absence of cost analysis creates major issues regarding accessibility and fairness within the cultural sector. Only large museums with substantial funding can afford high-cost systems, which increases resource inequality among cultural institutions. The Louvre Museum possesses sufficient funds to adopt Bluetooth-based tracking technology, whereas smaller regional museums lack this financial capability. The gap between institutions results in unequal personalized experiences because visitors to bigger museums get advanced technology while those at smaller venues receive standard tours. Future research needs to deliver detailed implementation and maintenance cost assessments for proposed systems while conducting cost–benefit analyses and investigating affordable alternatives that ensure scalability and affordability. Future research that resolves these current limitations will enable museum administrators to receive realistic and actionable insights, which will make innovative systems more accessible to numerous institutions. The proposed approach will make personalized experiences more achievable while boosting equity and inclusiveness within the cultural sphere.

Development of personalized content delivery systems in museums depends on solving challenges like limited generalizability and self-reported data use, along with technical limitations and privacy issues, plus scalability and usability obstacles. Researchers who direct their efforts towards these areas will develop stronger systems that scale effectively and improve user experience to boost visitor engagement and experiences. Further studies must investigate how emerging technologies like artificial intelligence and augmented reality can create more personalized and enriched museum experiences.

4.5. Practical Implementation Considerations for Museums

This section unifies key findings from the literature review to improve the practical application of personalization technologies in smart museums. The objective aims to help museum professionals choose suitable technologies while planning implementation strategies and assessing outcomes in relation to their institutional capacity and visitor context.

The adoption of personalization technologies in museums must match the institution’s scale and its available resources and digital advancement level. Smaller museums facing budgetary and technical constraints can start with affordable options like mobile applications with preset content routes along with survey-driven personalization tools. The implementation of these technologies demands minimal infrastructure, and they can be incorporated into current visitor flows with limited training and support requirements.

Medium-sized museums demonstrate greater adaptability in adopting interactive technologies like Bluetooth beacon location tracking systems and touch displays that react to visitor actions, along with recommendation engines for route suggestions. Personalization features provided by these technologies enable customization according to exhibition requirements or visitor demographic information.

Larger institutions benefit from their technical expertise and funding to implement advanced personalization systems. Institutions that can support advanced systems might implement artificial intelligence platforms for behavioral analysis as well as multimodal sensor networks that gather movement and emotional data while adaptive content delivery systems dynamically adjust to visitor preferences and interactions. The deployment of these technologies needs thorough planning and maintenance and ethical analysis but can create highly personalized visitor experiences.

The implementation must be adapted to fit both the museum type and its specific thematic subject matter. Art museums stand to gain from personalized content depth systems which adapt to visitor expertise and interest levels. Science and technology museums can boost visitor engagement through gamified personalization tools and interactive learning interfaces, which appeal especially to young museum visitors. Mobile guide systems with geolocation, augmented reality overlays, or virtual historical scene reconstructions work best at open-air and heritage sites because their displays require spatial and contextual understanding.

Visitor data collection and management becomes an essential area for implementation. Academic sources emphasize the significance of merging passive data like exhibit time and visitor movement with active feedback via surveys and mobile app push notifications. Data collection practices need to uphold privacy principles and ethical guidelines as their foremost priority. Data collection processes must involve clear visitor communication about data practices while obtaining needed consent and guaranteeing data anonymization as well as secure storage. Transparent system designs must enable users to understand and manage their data usage within their experience.

The assessment of personalization initiative outcomes is crucial to ensure enduring development and sustained performance. Useful engagement indicators can be derived from quantitative measures, including average dwell time alongside interaction count and revisit rates. Qualitative assessments of visitor satisfaction alongside emotional reactions and perceived educational benefits serve to enhance the analysis of engagement data. Organizations should pursue longitudinal research whenever feasible to monitor personalization effects on repeat visits and knowledge preservation, as well as wider community participation over time.

By integrating these practical considerations into planning and evaluation processes, museums can make informed decisions that align technological capabilities with their mission, audience, and operational context, ensuring that personalization efforts contribute meaningfully to both visitor experience and institutional goals.

4.6. Future Research Directions

Content personalization in smart museums shows significant progress in current research, but many important aspects need further investigation. Future research programs should focus on developing interdisciplinary approaches that link museum research with computer science and cognitive psychology. Specific research questions to be addressed include the following: How can personalization systems respect user privacy while offering meaningful adaptation? What cognitive and emotional mechanisms are triggered during personalized cultural experiences? How can content be adapted to culturally diverse audiences or visitors with different experience levels?

Emerging technologies such as brain–computer interfaces (BCIs) and emotion AI present new opportunities to dynamically adapt content based on real-time cognitive and affective states. Although still experimental, these technologies warrant further exploration in museum settings, particularly in enhancing inclusivity for neurodivergent visitors or those with cognitive impairments.

There is a significant gap in research focused on longitudinal evaluations of the lasting effects of personalization technologies. Follow-up research should explore how personalization technologies affect visitor knowledge retention and re-engagement, along with perceived cultural value over the long term. Institutional goals and individual experiences will only benefit from personalization if researchers create scalable methods that remain ethically sound.

5. Conclusions

The review systematically examines how smart museums apply different methods and technology to understand visitor behavior and create personalized experiences. Through an analysis of 33 studies performed from 2015 to 2024 that explore visitor engagement enhancement at museums, this review highlights the essential role of data-driven methods and contemporary technologies. Through the integration of Statistical and Data Analysis (SDA) and Artificial Intelligence and Machine Learning (AML) with Mobile and Interactive Technologies (MIT), museums gain accurate insight into visitor behavior patterns. Museums can build customized content and design responsive environments which meet visitors’ current needs by applying these methodologies.

This research determined Interactive Engagement and Movement Patterns to be the key behaviors assessed, which showed the importance of studying visitor interactions with museum exhibits and their navigation through the museum environment. Museums have successfully utilized Behavioral Analytics Platforms combined with Location Tracking Systems and AI/ML Systems for their exceptional capability to capture visitor behavior patterns, which enables museums to enhance exhibit placement and deliver targeted information to visitors. The research revealed shortcomings in Emotional Response and Group Dynamics studies, which indicate future research should combine affective computing technology with social interaction methods to create museum experiences that support empathy and collaboration.

While personalization technologies have achieved significant advancements, their development continues to face numerous unresolved problems. The adoption of advanced systems proves difficult for smaller museums due to ongoing privacy concerns combined with scalability issues and significant implementation costs. Current research encounters limitations in generalizability and long-term impact because it relies heavily on self-reported data without sufficient longitudinal study support. Future research needs to develop scalable and ethical solutions which offer affordable implementation methods across different museums.

The review demonstrates that interdisciplinary collaboration remains crucial for comprehending the intricate aspects of visitor behavior. Interdisciplinary researchers bring together computer science knowledge with psychological and educational insights along with cultural heritage expertise to develop visitor-focused personalization strategies. Museums can use Virtual/Augmented Reality and advanced biometric systems to develop immersive experiences which emotionally connect with visitors but need to maintain technological integration alongside ethical standards and visitor privacy protection.

The evaluation demonstrates how visitor behavior analysis together with personalization technologies can transform smart museums. The utilization of advanced technologies and data models by museums enables the creation of visitor experiences that connect with and accommodate all people. To harness these innovative advancements, museums need to solve existing problems and investigate new research prospects. Future research needs to develop scalable approaches that ethically target visitors to improve museum experiences by providing better accessibility and universal personalization benefits.