Content area
The research investigates the physical and psychological stress experienced by pilgrims during Hajj and Umrah through the implementation of advanced technologies which include wearable sensors and remote sensing systems. The study achieves a "comprehensive understanding" through its multidimensional approach which combines physiological data with environmental information and subjective feedback to analyze stress dynamics beyond previous pilgrimage research. The framework demonstrates strong methodological quality through its combination of real-time biometrics with geospatial tracking and validated questionnaires but its small sample size restricts generalization. The framework provides a fundamental model for detecting essential stress thresholds but future research should focus on expanding the system and conducting extensive validation tests. The system tracks heart rate and energy expenditure and fatigue levels through Apple Watches and mobile applications and web-based analytics software which combine these biometric measurements with environmental data points that include geographical positions and activity types. The research delivers practical findings about high-stress levels which guide the development of specific interventions to boost pilgrimage experiences despite its restricted participant pool. This research leads the way for innovative solutions which align with Vision 2030 to enhance crowd management and healthcare services and overall well-being during Hajj and Umrah.
Introduction
The Hajj and Umrah pilgrimages represent some of the largest religious gatherings globally, drawing millions of participants each year to fulfill their spiritual obligations. Although these experiences are of great religious significance, they expose pilgrims to considerable physical and psychological challenges. Factors such as long walking distances, extreme heat, crowded spaces, and time-sensitive rituals often result in elevated levels of stress and fatigue. Managing these challenges is critical to enhance the overall experience of pilgrims and ensure their health and safety throughout the journey. Effective stress measurement and analysis during Hajj and Umrah can provide valuable insights into addressing these issues, enabling better preparation, resource allocation, and intervention strategies to alleviate stress during the pilgrimage.
Despite the promising applications of remote sensing and machine learning, several challenges and limitations persist in implementing these systems in high-density pilgrimage settings. One significant challenge is the variability of individual responses to stress, which can complicate the development of generalized models (Gedam and Paul 2021; Kiranashree et al. 2021). Additionally, the integration of diverse data sources, physiological, environmental, and behavioral, requires sophisticated data management and analysis techniques, which can be resource-intensive (López et al. 2023). Another challenge is the ethical consideration of monitoring individuals without consent, particularly in a religious context where privacy is paramount. Ensuring that data collection methods are respectful and transparent is essential to gaining the trust of pilgrims and stakeholders (Gedam and Paul 2021; Kiranashree et al. 2021). Furthermore, the technical limitations of remote sensing systems, such as battery life and data transmission capabilities, can hinder their effectiveness in crowded environments (López et al. 2023).
Insights from similar studies in high-stress, high-density events research conducted in other high-stress, high-density events, such as concerts, sporting events, and natural disasters, provide valuable insights that can be applied to studying stress during Hajj and Umrah. For instance, studies have shown that environmental factors, such as noise levels and crowd density, significantly impact stress levels in concert settings (Dalmeida and Masala 2021; Gedam and Paul 2021). Similarly, machine learning applications in disaster response have demonstrated the effectiveness of real-time data analysis for predicting stress-related outcomes (Iqbal et al. 2022). These insights underscore the importance of interdisciplinary approaches that combine knowledge from various fields, including psychology, public health, and data science, to develop effective stress management strategies during pilgrimages. By leveraging lessons learned from other high-stress events, researchers can enhance the understanding of stress dynamics in the unique context of Hajj and Umrah.
The integration of remote sensing systems and machine learning in studying pilgrims’ stress during Hajj and Umrah is an emerging field with significant potential for future research. The research community should focus on developing more sophisticated models, such as data fusion models, that account for individual variability in stress responses and the dynamic nature of the pilgrimage environment. Additionally, exploring advanced machine learning techniques, such as reinforcement learning and transfer learning, could enhance the predictive capabilities of stress models (Fauzi et al. 2022). Moreover, there is a need for longitudinal studies that track stress levels over time, providing insights into the long-term effects of pilgrimage on mental health. Such studies could inform the development of targeted interventions to reduce stress and improve the overall pilgrimage experience for millions of participants (Gedam and Paul 2021; Kiranashree et al. 2021).
There are several gaps in the literature regarding the measurement and analysis of stress among pilgrims during Hajj and Umrah. One significant gap is the inability of current models to account for individual variability in stress responses. Given pilgrims’ diverse demographic and cultural backgrounds, research is needed to develop adaptive and personalized stress measurement models that accommodate these differences. Additionally, integrating multimodal data, such as physiological, environmental, and behavioral inputs, remains a technical challenge due to the complexity of managing and analyzing such datasets. Advancements in data fusion techniques and real-time analytics could enhance the accuracy of stress prediction models. Another notable gap is the lack of longitudinal studies that track stress levels over time during and after the pilgrimage. These studies could provide valuable insights into the long-term impacts of the pilgrimage experience on mental and physical well-being. Furthermore, ethical considerations surrounding data collection in religious contexts are underexplored. Issues of privacy, consent, and cultural sensitivities require more attention to ensure the respectful and transparent use of data collection technologies. Developing frameworks for ethical data collection could address these concerns and gain the trust of stakeholders.
Many studies on stress measurement are based on unrelated high-density events such as concerts or sports, with few explicitly focused on religious gatherings like Hajj and Umrah. Conducting context-specific research for stress analysis in these different environments could also provide more accurate and relevant insights. Addressing these gaps requires an interdisciplinary approach that leverages advancements in wearable technology, machine learning, ethics, and cultural studies to develop scalable, personalized, and ethically sound solutions tailored to the challenges of Hajj and Umrah.
The Hajj and Umrah pilgrimages draw millions of participants yearly because they combine large numbers of people with harsh environmental conditions and deep spiritual engagement. The management of stress in these environments remains essential to achieve better safety outcomes and improved comfort and total experience for pilgrims. The current research contains multiple unaddressed knowledge gaps. The current research lacks specific data about Hajj that combines biometric measurements with geographic information to understand the distinctive challenges of this pilgrimage. In Alharthi and Alshamarani (2021) and Al-Shaery et al. (2022) along with other studies have mainly used survey methods to gather data which provides valuable perception data but does not fully measure real-time physiological responses of pilgrims. The absence of real-time multimodal stress monitoring systems designed for pilgrimage settings prevents the delivery of immediate support and intervention measures. The research establishes a complete system which uses wearable sensor information together with environmental and behavioral elements to track and evaluate real-time stress levels during Hajj and Umrah. The research applies advanced technologies to generate practical knowledge which enhances crowd management and healthcare services and pilgrim wellness according to Vision 2030 goals.
This study addresses a critical research problem: the lack of real-time, data-driven insights into the physical and emotional stress experienced by pilgrims. Existing studies have primarily relied on subjective surveys or isolated physiological data, which fail to capture the dynamic and multifaceted nature of stress during large-scale pilgrimage events. Furthermore, there is limited integration of advanced technologies, such as wearable devices and remote sensing systems, to analyze stress and identify contributing factors. Recognizing this gap, the study aims to provide an evidence-based approach to stress management using modern technologies.
The objectives of this study are twofold: first, to collect and analyze real-time physiological and spatiotemporal data using wearable devices such as Apple Watches; second, to identify critical factors, locations, and activities that contribute to stress and tiredness during Hajj. By achieving these objectives, the study aims to provide actionable insights for managing stress and improving pilgrims’ well-being during their journey. To guide this research, the following research questions were posed:
What are the primary physiological and environmental factors contributing to stress and fatigue during Hajj and Umrah?
How do stress levels vary across specific locations, time periods, and pilgrimage activities?
We propose a comprehensive mixed-methods framework for collecting data for Hajj and Umrah targetting stress analysis and detection.
We implement our proposed framework into a fully functional system that is used in a real-world case study.
We present a case study utilizing our proposed system, conducted during the Hajj of the year 1445 AH (2024 AD), which collected real-world data.
We thoroughly analyze the collected data and discuss insightful findings.
Related work
Recent advancements in remote sensing systems and machine learning offer promising avenues for measuring and analyzing stress in such dynamic, large-scale environments. This section discusses related work, including methodologies, technologies, and applications. In addition, we discuss challenges and limitations related to high-density religious pilgrimage settings.
Stress Measuring Methods
Traditional methods for measuring stress often rely on self-reported questionnaires and physiological indicators such as heart rate variability (HRV), galvanic skin response (GSR), and cortisol levels. However, these methods can be limited by their reliance on subjective reporting and the need for controlled environments. Recent studies have highlighted the potential of wearable devices and remote sensing technologies to provide real-time, objective measurements of stress indicators. For instance, wearable sensors can continuously monitor physiological parameters, enabling a more nuanced understanding of stress responses in real-time (Dalmeida and Masala 2021; Gedam and Paul 2021). Moreover, the integration of remote sensing systems, such as drones and satellite imagery, has been explored for monitoring large crowds during events like Hajj. These systems can capture environmental data, including temperature, humidity, and crowd density, which can correlate with stress levels among pilgrims (López et al. 2023). The combination of physiological data from wearables and environmental data from remote sensing can provide a comprehensive picture of the stress experienced by individuals in such high-density settings.
Remote sensing systems for stress monitoring
Remote sensing systems are crucial in monitoring both physiological and psychological stress during pilgrimages. These systems can collect vast amounts of data that can be analyzed to identify stress patterns among large groups. For example, studies have utilized drone technology to assess crowd dynamics and environmental conditions, which are critical factors influencing stress levels (López et al. 2023). Monitoring these variables in real-time allows for timely interventions and better crowd-management strategies. Additionally, remote sensing can be integrated with physiological monitoring systems to create a more holistic approach to stress assessment. For instance, combining heart rate data from wearables with environmental data from remote sensors can help identify specific stress triggers during the pilgrimage (Dalmeida and Masala 2021; Gedam and Paul 2021). This integrative approach enhances the accuracy of stress measurement and provides insights into the interactions between environmental factors and individual stress responses.
Stress detection using wearable devices in everyday scenarios was also explored in alignment with the research presented in this study (Can et al. 2019). Data for these studies were collected from the WESAD database, which includes a combination of physiological and motion datasets. This data incorporates sensor readings such as electroencephalogram (EEG), electrocardiogram (ECG), and motion sensors to capture a person’s movement (Schmidt et al. 2018). The data was recorded using wrist-worn devices and chest-mounted sensors designed to identify various behavioral and interactive states (neutral, stressed, and amused) in alignment with the objectives of this study. This dataset is utilized to predict an individual’s stress levels by analyzing motion and sensor-based biometric readings.
Although the WESAD dataset includes stress detection data, it lacks geographical information about the locations where the data was collected. In fact, one significant factor that may influence a person’s stress level is their current location, which is particularly relevant to our study. For example, being in a crowded train station could heighten a person’s stress level. Similarly, knowing whether someone is walking in a large plaza, or a busy main street could enhance the accuracy of stress level predictions.
Unlike the studies, we aim to collect a stress dataset specific to the Hajj and Umrah contexts to identify critical factors, locations, and activities that contribute to stress and tiredness during Hajj. To the best of our knowledge, no publicly available dataset for researchers combines health data from pilgrims (e.g., heart rate, blood oxygen levels, stress levels, and fainting incidents) with location information. This gap represents a significant limitation in current research, which our study aims to address.
Stress detection among pilgrims, especially during religious activities such as Hajj and Umrah, has gained significant attention in recent research due to the potential health risks posed by stress. Several studies have been conducted to analyze the stress levels and characteristics of pilgrims using advanced technologies, including machine learning and big data analytics. Alharthi and Alshamarani (2021) conducted a study on pilgrims residing in Makkah, using machine learning algorithms to analyze their characteristics and detect potential stress levels. The research focused on monitoring pilgrims’ physiological data, behavior patterns, and environmental factors that could contribute to stress during their stay in Makkah, providing insight into how machine learning can help in real-time monitoring and stress detection. The study emphasized the importance of using advanced data analytics to support the management of large crowds and ensure the safety and well-being of pilgrims during peak times.
In a similar vein, Elgamal and Alshamarani (2019) presented an analytical study that explored the key characteristics of the guests of Al-Rahman, using modern technologies to mine large datasets. This study aimed to provide deeper insights into the pilgrims’ behavior and stress triggers by leveraging data analytics. The findings suggested that identifying specific stress-related patterns, such as physical exhaustion, overcrowding, or environmental stressors, can help develop better support systems for pilgrims.
A crowd control challenge during Hajj was explored in a study through the development of an expert system which integrates wearable physiological sensors, GPS, and deep learning. An Android apps and devices like Zephyh BioHarness and Empatica E4 are used to monitor the physiological parameters and movements of the pilgrims in real-time. A bidirectional LSTM model could recognize Hajj rituals with 75% accuracy and fatigue with 92% accuracy, but emotion classification was low because of its continuous scale measurement. The results of the present research with a limited sample size show that sensor fusion and AI can be useful for improving crowd safety, specifically for hazards like stampedes. One gap in the research has been addressed by the authors through the inclusion of physiological data in addition to the traditional spatial and visual crowd analysis (Al-Shaery et al. 2022). A survey on IoT and AI based remote healthcare monitoring systems use wearable sensors and machine learning techniques such as SVM and k-means for real time analysis of health parameters including heart rate and body temperature (Alshamrani 2022). The study also identifies IoT protocols such as Bluetooth and Zigbee for efficient data transmission and anomaly detection which is directly applicable to frameworks that are monitoring stress and fatigue in Hajj pilgrims. Problems such as energy efficiency and data reliability in crowded settings are consistent with the requirements of real-time crowd management, and require scalable IoT architectures to provide accurate health information during mass gatherings (Alshamrani 2022).
Wearable sensors for monitoring stress and fatigue have emerged as an important area of interest in recent years due to the advancements in technology and the need for continuous monitoring of health parameters. Today, wearable devices are able to capture subtle changes in the physiological and psychological state of an individual and thus can be used in different areas, including workplace safety, sports performance, and health care. Fatigue can be measured through several physiological parameters by wearable sensors. For example, research has revealed that wearable devices are capable of tracking heart rate variability which is a measure of fatigue. In particular, Bae et al. pointed out that the real-time heart rate monitoring during the CPR can signify the fatigue of the providers which indicates that such devices can be useful in stressful situations in a specific environment (Bae et al. 2021). Similarly, Chan et al. reviewed the feasibility of wearable IMUs in the assessment of fatigue-related changes in spine motion, which may prevent musculoskeletal disorders in the workplace and sport (Chan et al. 2020). These findings are significant as they show that wearable technology can help in timely identification of fatigue, which can improve performance and safety.
Furthermore, the creation of new materials and sensor technologies has led to the improvement of the wearable devices. In this regard, Xiao et al. explained how electrospun fibers can be used in the fabrication of flexible strain sensors that can be incorporated into clothing for the overall monitoring of body movements and physiological functions (Gao et al. 2023). This integration enables the capture of various data types such as sound, pulse and movement which can be transmitted at real time to monitoring platforms for further analysis. The possibility to obtain multimodal data increases the accuracy of the fatigue diagnosis and gives a more comprehensive picture of a person’s health.
The use of machine learning algorithms in conjunction with wearable sensors greatly improves their performance in the detection of fatigue. For example, Zeng et al. proposed the epidermal electronic system teamed up with machine learning to unobtrusively detect mental fatigue and SPGM model was used to achieve accurate fatigue classification based on physiological signals (Al-Shaery et al. 2022). Also, Shi et al. proposed a gait pattern based fatigue recognition system to identify the level of fatigue and the results showed that it is possible to design wearable devices that can help in the assessment of an individual’s fatigue status (Shi et al. 2022). These technological enhancements also enhance the credibility of fatigue assessments and help in individualized approaches to health care. Wearable devices also have potential in helping individuals with chronic conditions to cope with fatigue in specific populations. For instance, Pereira reported that the utilization of wearable devices enhanced the quality of life and decreased the fatigue in patients with inflammatory myopathies (Pereira et al. 2024). This shows that wearables can be useful in the management of health care delivery to special groups of patients. Therefore, wearable sensors can be viewed as revolutionary technology for stress and fatigue management in real-time. Their suitability for continuous, non-contact monitoring via improved materials, machine learning and multimodal data analysis makes them valuable assets in the field of health care across different fields. With the rate at which technology is evolving, there are likely to be more and more ways in which wearable devices can be used in the management of fatigue and other aspects of health, which may go a long way in improving the quality of life of the individual.
[See PDF for image]
Fig. 1
Our comprehensive mixed-methods framework for collecting and analyzing the data for our study
Wearable sensors have gained significant traction in the realm of stress detection and data analysis due to their potential for continuous physiological monitoring and real-time feedback mechanisms. These devices leverage various physiological signals, including heart rate variability (HRV), skin conductance, and body temperature, to assess an individual’s stress levels. The integration of machine learning algorithms with these signals has further advanced the capability of wearables in accurately detecting stress. Continuous monitoring is crucial for effective stress management, as it enables early detection of stress-related issues before they escalate into more severe health problems. A comprehensive review indicated that wearable devices facilitate real-time and continuous data collection pertinent to personal stress monitoring, thereby enhancing individuals’ ability to manage stress proactively (Gedam and Paul 2021). Furthermore, physiological signals such as Galvanic Skin Response (GSR) and skin temperature have been used to classify stress-related events, as they are influenced by the autonomic nervous system (Ehrhart et al. 2022). This supports the notion that physiological measurements can significantly correlate with stress levels, providing necessary data for real-time intervention (Alam and Alam 2023). Recent advancements have shown the efficacy of deploying multimodal sensor systems for enhanced stress detection. For instance, a study highlighted the effectiveness of combining physiological measurements, such as heart rate variability and contextual data, to track stress progression accurately (Talaat and El-Balka 2023). This integration allows for a more nuanced understanding of stress triggers and responses, paving the way for personalized stress management strategies. Moreover, innovative approaches are being explored to improve the accuracy of stress detection through wearable sensors. A conditional generative adversarial network (GAN) has been developed that generates time-series data for stress detection, showing that such data can enhance predictive models (Ehrhart et al. 2022). Additionally, semi-supervised learning techniques are being employed to attain higher accuracy rates in detecting momentary stress in dynamic and uncontrolled environments (Yu and Sano 2022). Such machine learning models effectively utilize both physiological data and behavioral indicators to predict stress in real time. In terms of application contexts, the use of wearable sensors is notably impactful in medical settings, especially during pandemics where stress levels may intensify due to external pressures. For example, the utilization of wearable technology to monitor healthcare workers’ stress levels during crisis situations, such as the COVID-19 pandemic, highlights the importance of real-time physiological data to support mental health (Hosseini et al. 2022). Collaborative efforts that combine contextual information (for example, location, activity) with physiological signals indicate a promising synergy for future wearable technologies (Rashid et al. 2023). Lastly, research continues to advocate for extensive studies that integrate multiple sensor types and improve machine learning algorithms designed for stress detection. Such advancements can not only refine stress monitoring accuracy but also provide personalized feedback and interventions suited for individual users, marking a significant step forward in health technology (Pote et al. 2024; Adam et al. 2024).
Hajj-specific stressors
The Hajj pilgrimage creates distinctive stress factors which separate it from concerts and sports events because of its cultural and religious elements and operational requirements. The religious and cultural nature of Hajj creates a spiritual journey that requires pilgrims to complete specific rituals during specific time periods at designated locations while facing intense time constraints among large crowds. The deeply spiritual nature of the experience creates increased emotional and psychological stress because pilgrims perform their religious duties in an intense environment. The extreme heat exceeding 40°C in the environment creates major physical difficulties for pilgrims especially those who are elderly or have health issues. The diverse languages spoken by pilgrims create communication barriers which may cause misunderstandings and anxiety among participants. The massive number of participants who gather in a small space during Hajj creates extremely dense crowds which intensifies stress levels and increases safety risks. The distinct elements of Hajj require specialized stress monitoring and management strategies which demonstrate why research and interventions must be developed specifically for this pilgrimage.
Research methodology
This study uses a comprehensive mixed-methods approach to assess and analyze stress levels among various Hajj and Umrah pilgrims. Participants are selected through stratified sampling techniques to ensure a diverse representation across middle age groups (35-45 years) and genders. We ensured that all participants were in good health. The study design captures data throughout the pilgrimage journey, including pre-departure, transit, and on-site rituals. This provides temporal and spatially relevant data, offering a deeper understanding of stress dynamics during different stages of the pilgrimage. Figure 1 depicts our methodology for data collection.
To facilitate the data collection process in this study, we built a system that comprises an interconnected ecosystem of hardware and software tools tailored to capture, manage, and analyze multidimensional stress-related data. The framework includes the following:
Apple watch app:
Watch App: A custom-designed application functions as the primary tool for real-time physiological monitoring. It continuously tracks key stress indicators, including:
Heart rate and variability (HRV), which reflect autonomic nervous system activity.
Step count and distance traveled, providing insight into physical exertion.
iPhone app:
Complementing the Apple Watch, the iPhone application extends the system’s capabilities by providing the following features:
Interactive Logs: Logs participants’ subjective experiences, such as perceived stress levels or challenges encountered during rituals.
Data Visualization: Displays stress trends intuitively, helping users track their physiological responses over time.
Integration with Environmental Data: Synchronizes with location-based inputs, such as GPS coordinates, to overlay physiological data with environmental conditions, including crowd density and ambient temperature.
Personalized Feedback: Offers real-time stress management tips tailored to individual data patterns.
Web-based admin panel:
We also designed a web application to assist researchers and administrators in managing and analyzing study data.The admin panel serves as a centralized hub with key features, including:
Participant Tracking: Monitors participant status, tracking data submission progress and identifying inconsistencies or missing data.
Off-line Dashboards: Provides interactive visualizations of aggregated data, including stress records and demographic-specific trends.
Advanced Analytics Tools: Supports complex data analyses, such as clustering.
Scalable Infrastructure: Built with cloud-based architecture to handle datasets efficiently while ensuring robust security protocols.
Participants questionnaires:
A series of custom questionnaires is developed to complement physiological data collection. These questionnaires are structured into three categories:
During performing Surveys: Administered through the iPhone app, these capture real-time contextual data during rituals, such as perceived challenges, emotional states, and environmental stressors.
After performing Surveys: Focus on retrospective insights, assessing participants’ overall experience, satisfaction, and recovery patterns. Each questionnaire undergoes validation and testing to ensure clarity, reliability, and relevance. Questions are designed to elicit specific data aligned with the study’s objectives.
Data collection process and analysis:
The data collection process integrates physiological data (e.g., heart rate, blood pressure, and blood oxygen levels), physical activity data (e.g., distance walked and number of steps), and self-reported data into a unified dataset. Key steps include:
Wearable Sensors: Apple Watch devices collect physiological data and automatically sync it with the iPhone app.
Location Tracking: GPS data records participants’ movements, allowing correlation of stress markers with specific pilgrimage locations (e.g., Tawaf, Sa’ai).
Time-based Sampling: Data is collected during peak and off-peak pilgrimage times to capture a comprehensive range of stress dynamics.
Ethical considerations:
Ethical compliance is integral to the study’s framework, ensuring participant rights and data protection by addressing key concerns as follows:
Informed Consent: Participants receive comprehensive information information about the study’s purpose, procedures, and data usage. Consent is obtained in a culturally sensitive manner, ensuring participants’ understanding and voluntary participation.
Data Privacy and Security: Anonymization protocols and encryption mechanisms protect participants’ datain compliance with international research ethics standards.
Cultural Sensitivities: The system design and implementation are aligned with the spiritual and cultural context of Hajj and Umrah, ensuring respect for religious practices.
Transparency: Participants can access their data and insights, fostering trust and engagement.
System implementation
Implementing the system marks a critical phase in transforming the conceptual framework into an operational platform capable of collecting, analyzing, and managing stress-related data from pilgrims during Umrah and Hajj. This phase integrates software applications, hardware components, and data management tools into a unified system to deliver accurate and actionable insights. The implementation focuses on ensuring reliability, user-friendliness, and scalability to accommodate diverse pilgrimage activities and conditions. Key aspects of the process include deploying and optimizing the Apple Watch and iPhone applications for performance and battery efficiency, establishing robust communication channels between the apps and the web-based admin panel, and creating a seamless mechanism for data integration and synchronization. Emphasis is placed on iterative testing and validation to refine the system’s performance, addressing any technical or usability challenges during development. The following sections outline the deployment of the applications, testing and verification processes, and the strategies used for effective data integration and synchronization.
Deployment of the apps
The deployment of the Apple Watch and iPhone applications required careful planning to ensure seamless functionality and user accessibility. The applications were developed using Swift, a modern programming language, valued for its safety, speed, and efficiency, to build the apps. Integrating the HealthKit framework allowed the application to collect, store, and share physiological metrics across devices, such as heart rate and blood oxygen levels. The Core Data framework managed and stored complex datasets locally, ensuring reliable data handling and offline access. The deployment process involved the following stages:
Initial Rollout: The applications were deployed in an internal testing environment to identify and resolve technical issues, including device synchronization delays and intermittent data transfer problems.
User Onboarding: A streamlined onboarding process helped participants quickly understand and start using the apps. Automated prompts within the iPhone app guided users through setup and initial usage.
Technical Challenges: Significant challenges during deployment included differences in hardware performance across Apple Watch models and compatibility issues with various iOS versions. Developers conducted extensive device testing to identify and resolve these problems, optimizing the applications based on test results to ensure stability and performance.
Testing and validation
The system underwent a comprehensive testing and validation process to ensure its robustness, usability, and accuracy:
Functionality Testing: Both apps underwent extensive unit and integration tests to confirm real-time data collection, user interactions, and environmental data synchronization.
Pilot Study: A pilot study was conducted with a representative sample of participants to evaluate the system’s real-world performance. Feedback was collected on usability, battery efficiency, and data accuracy.
Iterative Refinements: Based on participant feedback and observed performance metrics, multiple iterations were implemented to address identified issues. These included improving the accuracy of step counts, optimizing app interfaces for user-friendliness, and enhancing the reliability of data transmission.
Validation Metrics: Physiological readings from the Apple Watch were cross-referenced with clinical-grade devices to validate data accuracy. Additionally, user-reported experiences were analyzed to align perceived stress indicators with collected data.
Table 1. Description of features and health metrics collected by smartwatch
Feature | Description |
|---|---|
First Name | The participant’s first name |
Last Name | The participant’s last name |
The participant’s email address | |
Gender | The gender of the participant |
Category Name | The geographic place assigned to the participant |
Zone Name | The geographic zone in the place assigned to the participant |
Date Time | The date and time when the data was recorded |
Heart Rate | The participant’s heart rate in beats per minute as measured by the smartwatch |
Steps | The number of steps the participant took during a given time period |
Resting Energy | The amount of energy in calories burned while the participant was at rest |
Active Energy | The amount of energy in calories burned during physical activities |
Walking Distance | The total distance in miles covered by the participant during walking or running activities |
Walking Asymmetry | The asymmetry in the participant’s walking pattern, referring to an imbalance in the way a person moves, |
specifically differences between the left and right sides of the body during walking | |
Walking Step Length | The average length of the participant’s steps in centimeters |
Walking Speed | The participant’s walking speed in kilometers per hour |
Double Support Time | The time in seconds or milliseconds both feet were on the ground during walking |
Standing Minutes | The number of minutes the participant spent standing during a given time period |
Blood Oxygen | The participant’s blood oxygen saturation level measured in a percentage |
Body Weight | The participant’s body weight in kilograms |
Body Height | The participant’s body height in centimeters |
Data integration and synchronization
Integrating data from multiple sources formed a cornerstone of the system’s functionality, enabling a cohesive analysis of physiological and contextual data. The integration process involved:
Centralized Data Repository: A secure cloud-based repository was established to aggregate data from the Apple Watch, iPhone app, and web-based admin panel, using scalable database architecture to handle high data volumes.
Real-Time Synchronization: The Apple Watch app transmitted physiological and physical data using Bluetooth Low Energy (BLE) to the iPhone app. The iPhone app uploaded this data, user-input logs, and environmental context to the cloud repository in near real-time.
Data Standardization: The system preprocessed collected data to standardize formats across sources. Timestamp synchronization aligned of physiological readings with corresponding environmental metrics.
Web-Based Panel Integration: The admin panel accessed the centralized repository through secure APIs, providing researchers with tools to visualize and analyze integrated datasets. Advanced filtering options enabled segmentation by participant demographics, location, and activity phase.
[See PDF for image]
Fig. 2
Sample of health metrics collected by the Apple Watch
Stress data collection for Hajj 1445 season
The proposed dataset is designed to study the heat stress levels of Hajj participants during the pilgrimage. Stress data covers five days of Hajj 1445 AH (2024 AD), from June 13 (the 7th of Hajj) to June 18 (the 14th of Hajj). Three participants (two women and one man, aged between 30 and 50) used Apple Watches (Ultra 2 and Series 9 models) to collect data. The participants included two women and one man, aged between 30 and 50.
The Apple Watches captured participant data, including personal information (first name, last name, email, gender), zone, date, time, and health metrics. Health metrics included heart rate, steps, walking parameters (e.g., distance, speed, asymmetry), standing time, blood oxygen, body weight, and height. Table 1 summarizes each feature recorded by the smartwatches.The study also includes a twelve-question questionnaire to gather qualitative data on participants’ levels of tiredness and emotional states throughout their Hajj journey. Appendix A provides a detailed description of each question.
Data cleansing
In this section explains how to handle missing data and integrate health information with questionnaire responses to form the final dataset. Figure 2 presents a sample of health metrics collected by the Apple Watch, showcasing the key physiological data monitored throughout the study. The metrics include heart rate, step count, active energy expenditure, and walking distance. The metrics continuously recorded participants’ physical activity and health status, providing real-time insights into participants’ physical activity and health status during the Hajj pilgrimage. They helped identify patterns of fatigue and stress, particularly during physically demanding rituals. Heart rate data indicated periods of elevated exertion, while step count and walking distance reflected movement intensity, often surpassing 10,000 steps in a single day. Active energy expenditure, measured in kilocalories (kcal), quantified the physical demands placed on participants. Integrating these measurements with self-reported tiredness levels allowed the study to correlate objective health data with participants’ experiences, offering a clearer understanding of the physiological challenges during Hajj.
Handling missing data
This study collected health data from three participants during the Hajj pilgrimage. Each dataset underwent individual preprocessing to prevent inaccuracies from combined processing. Table 1 describes the features and health metrics collected by the smartwatch. To improve dataset quality, the preprocessing steps removed columns with more than 96% missing values. The preprocessing filled gaps using the most frequent values for features with moderate missing data, such as body weight, body height, category name, and zone. When feasible, this imputation used cross-referenced data from other participants. Time-series health metrics used forward and backward filling to interpolate missing values based on surrounding entries in the timeline. The preprocessing also excluded records collected outside the designated pilgrimage period to maintain data consistency and relevance. For gaps that sequential filling could not resolve, the k-Nearest Neighbors (KNN) method (K = 5) estimated missing values based on similar entries in the feature space. This individualized preprocessing approach refined each dataset, significantly improving data quality by effectively handling missing information.
Combining health information and questionnaire data
Integrating health information from wearable devices and subjective questionnaire responses provided a comprehensive and multidimensional understanding of stress and tiredness among pilgrims during Hajj. Combining real-time physiological data, such as heart rate, step count, active energy expenditure, and walking distances, with self-reported experiences revealed strong correlations between physical exertion and perceived fatigue or stress levels. Wearable devices like the Apple Watch Ultra 2 and Series 9 recorded continuous physiological metrics, providing objective insights into the participants’ health status during different stages of the pilgrimage. For example, heart rate spikes, significant increases in step counts (exceeding 10,000 steps), and active energy expenditure (above 450 kcal) are closely aligned with participants’ reports of “Tired” and “Extremely Tired” states, particularly on the 10th day of Hajj (Day 4). This correlation highlighted the cumulative nature of fatigue resulting from extended physical activity and environmental stressors, such as high temperatures and crowd density.
The questionnaire responses added subjective context to the quantitative health data, reporting emotional states, perceived tiredness, and challenges in specific locations (e.g., walking in Mina, overnight stay in Muzdalifah). These responses helped identify the spatial-temporal variations in stress, where locations like Mina and Muzdalifah emerged as the most physically demanding areas. At the same time, Arafah showed lower reported stress levels due to reduced physical exertion. Integrating both datasets enabled the study to pinpoint specific stress-inducing thresholds, such as walking distances over 5,000 meters or burning high energy levels, correlated with higher reported stress and tiredness.
The analysis also highlighted individual variability in stress responses based on age, gender, and fitness, emphasizing the need for personalized stress management. This combined analysis grounded the findings in data and real-world experiences, making the results quantifiable and relatable. Aligning objective health data with subjective perceptions captured a holistic picture of pilgrims’ stress and fatigue during the Hajj.
Hajj stress dataset
This section presents the final, cleaned dataset developed to analyze the stress among Hajj participants during the pilgrimage. Data collection occurred during Hajj 1445 AH (2024 AD) over five days, from June 13, 2024 (7th of Hajj) to June 18, 2024 (14th of Hajj). Three participants (two women and one man aged between 30 and 50) used smartwatches (Apple Watch Ultra 2 and Series 9) to collect quantitative health data. A customized questionnaire gathered qualitative data on the participants’ tiredness and stress levels throughout the journey.
The dataset consists of 120 records with 24 features, classified into health metrics, contextual data, personal details, and subjective responses. Health metrics (e.g., heart rate, steps, active energy, walking distance, walking asymmetry, walking step length, walking speed, body weight, body height). Contextual data features consist of location data (e.g., latitude, longitude, stage), time information (e.g., date and time of day), and environmental factors (e.g., temperature). Personal details include identifiers (e.g., first name, last name, and gender). Subjective responses cover emotional states (e.g., feelings and reasons for feelings) and tiredness levels (e.g., tiredness and reasons for tiredness). Figure 3 provides a sample of the Hajj stress dataset. The dataset consists of 120 records across 24 features, including first name, last name, gender, location, stage, latitude, longitude, heart rate, steps, resting energy, active energy, walking distance, walking asymmetry, walking step length, walking speed, body weight, body height, date, time of day, temperature, feelings, reason of feelings, tired, and reason of tired. Figure 3 shows a sample of the Hajj stress dataset.
[See PDF for image]
Fig. 3
A sample of the final Hajj stress dataset
[See PDF for image]
Fig. 4
Heatmap addressing the correlation between features and target variables
[See PDF for image]
Fig. 5
The relationship between tiredness levels and the highly correlated features for each day of the Hajj trip
Data analysis and discussion of results
This section explores and extracts valuable insights from the provided data. These insights help form assumptions and identify correlations between certain features, shedding light on how they may impact tiredness and stress levels. Figure 4 represents a heatmap that displays the correlation between the features and the target variables. The target variables, "Tiredness" and "Stress", strongly correlate with four specific features: Steps, Active Energy, Walking Step Length, and Walking Running Distance. In contrast, the other features exhibit correlation values below 0.5 with the target variables, indicating weaker relationships.
[See PDF for image]
Fig. 6
Tiredness levels across each day of the pilgrimage
[See PDF for image]
Fig. 7
Tiredness levels across each location in the pilgrimage
Figure 5 illustrates the relationship between tiredness levels and the highly correlated features throughout each day of the Hajj trip. The Walking running distance feature in Fig. 5 shows a significant elevation in tiredness levels when it exceeds 5,000 meters, with participants transitioning from "Not Tired" to "Slightly Tired" or "Very Tired." This pattern emphasizes the exponential increase in tiredness associated with longer walking distances. On the other hand, the Walking Step Length feature in Fig. 5 does not demonstrate a clear relationship with tiredness levels. Most records indicating "Slightly Tired" or "Very Tired" have step length readings similar to those classified as "Not Tired". However, on the 10th of Hajj (Day 4), participants with step lengths exceeding 0.59 meters reported feeling "Extremely Tired". The Active Energy feature in Fig. 5 reveals a proportional relationship with tiredness levels. Participants typically report feeling "Slightly Tired" when their active energy expenditure falls within the 300 to 450 kcal range. However, "Very Tired" and "Extremely Tired" levels become more prevalent when energy expenditure exceeds this range. Lastly, the Steps feature in Fig. 5 also exhibits a proportional increase in tiredness levels. Participants who exceed 10,000 steps begin to report "Very Tired" levels, with "Extremely Tired" recorded on the 10th of Hajj (Day 4).
Tiredness levels throughout the pilgrimage indicate that the 10th of Hajj (Day 4) recorded the highest level of tiredness, with participants frequently reporting feeling "Extremely Tired". In contrast, other days, such as the 7th of Hajj (Day 1), showed less severe exhaustion levels, with participants reporting a mix of stress levels, including "Very Good" and "Good" levels, though one account noted "Bad". By the 9th of Hajj (Day 3), participants began to report feeling "Slightly Tired" toward the end of the day. This gradual buildup of fatigue likely contributed to the increased tiredness observed on Day 4, when participants frequently reported feeling "Extremely Tired", as illustrated in Fig. 6.
[See PDF for image]
Fig. 8
Tiredness levels across each stage in the pilgrimage journey
[See PDF for image]
Fig. 9
Common reasons cited for tiredness among participants based on tiredness level
Tiredness levels varied across pilgrimage locations, as depicted in Fig. 7, with Mina and Muzdalifah reporting the highest levels of exhaustion. Participants indicated they felt "Extremely Tired" in these areas. In contrast, Arafah’s maximum reported fatigue level was "Slightly Tired." Within the Haram, a range of tiredness levels was observed, with "Very Tired" being the most commonly reported category. This pattern suggests that the physical demands of the pilgrimage, particularly in locations like Mina and Muzdalifah, significantly contribute to increased fatigue among participants, with exhaustion peaking on the 10th of Hajj (Day 4).
The relationship between tiredness levels and the stages of the pilgrimage, as illustrated in Fig. 8, shows that participants experienced a range of tiredness levels during the Tawaf and Sa’i stages, with "Very Tired" being the highest recorded. However, upon reaching Muzdalifah and during the stoning ritual in Mina, the highest level of tiredness observed was "Extremely Tired". The reasons for varying levels of tiredness among participants provide valuable insights into the factors contributing to their fatigue.
Figure 9 details the common reasons cited for tiredness. For participants reporting "Not Tired", the most common reason cited was the lower crowd density. However, many still reported adverse conditions, such as hot weather, very crowded areas, and feelings of thirst, although these factors did not significantly impact their energy levels. In contrast, participants who reported feeling "Slightly Tired" often attributed their tiredness to fatigue and sleepiness, with hot weather being another notable factor. Thirst, hunger, and missed meals during bus rides were also cited with similar frequency, suggesting that these environmental and physical factors had a more noticeable impact on their tiredness.
[See PDF for image]
Fig. 10
Common Reasons for Tiredness by Location
[See PDF for image]
Fig. 11
Correlation between the time of day, locations, and tiredness levels
The reasons cited by participants experiencing "Very Tired" levels were more pronounced. Increased physical activity and the stress of managing time-sensitive rituals led to a significant sense of tiredness. Commonly reported factors included hot weather, thirst, fatigue, and sleepiness. Crowded conditions further exacerbated their tiredness, with participants identifying congested areas as a significant source of discomfort. Anxiety was also prevalent, as participants expressed that getting separated from their groups heightened their stress and tiredness.
Participants who reported feeling "Extremely Tired" identified several significant contributing factors. Locations such as Mina and Muzdalifah, particularly in the evening, were linked to frequent reports of extreme fatigue. The primary reasons included hot weather, tiredness and sleepiness, and overcrowded conditions. Many participants also mentioned thirst and hunger, which further intensified their exhaustion. Additionally, the need to stay overnight in Muzdalifah contributed to their tiredness, as insufficient rest significantly increased their fatigue.
[See PDF for image]
Fig. 12
Average values of features correlated with the ’Not Tired’ tiredness level across locations, specifically Haram, Mina, Muzdalifah and Arafah
[See PDF for image]
Fig. 13
Average values of key features correlated with the ’Slightly Tired’ level across Haram, Arafah, and Muzdalifah
Figure 10 illustrates how participants’ experiences varied by location. In Haram, the primary discomfort reported was hot weather, though many noted that the area was relatively uncrowded. However, a few participants mentioned overcrowding and reported feelings of tiredness, sleepiness, and thirst. Some also expressed concerns about losing track of their companions, which added to their stress and discomfort during the visit. At Arafah, the intense heat was the only reason reported for increased tiredness, significantly affecting participants’ comfort levels. In Mina, participants predominantly experienced fatigue due to hot weather, the most frequently reported cause. Other contributing factors included overcrowding, sleepiness, thirst, and hunger. Muzdalifah emerged as one of the most exhausting locations, with participants experiencing the highest tiredness levels. Contributing factors included hot weather, extreme fatigue from the overnight stay due to uncomfortable conditions and lack of sleep, lack of meals during bus rides, and crowded conditions. Participants consistently reported feeling tired, sleepy, and thirsty during their stay there.
[See PDF for image]
Fig. 14
Average values of key features correlated with the ’Very Tired’ level across Haram
[See PDF for image]
Fig. 15
Average values of key features correlated with the "Extremely Tired" level across Mina and Muzdalifah
[See PDF for image]
Fig. 16
Average values of key features correlated with the ’Extremely Tired’ level across Mina and Muzdalifah
Figure 11 reveals significant patterns by illustrating the correlation between tiredness levels, time of day, and locations. By analyzing how participants’ tiredness fluctuates at different times and locations, we can identify whether specific periods or places contribute to increased fatigue. This correlation highlights the influence of environmental and temporal factors on tiredness. The figure shows that a high proportion of participants report feeling "Slightly Tired" during the afternoon at Arafah. In contrast, Haram records varying tiredness levels, ranging from "Slightly Tired" to "Very Tired" throughout most of the day. In the evening, Mina and Muzdalifah exhibit the highest fatigue levels, with many participants reporting extreme tiredness. These findings suggest that environmental factors, such as location, significantly impact tiredness levels more than temporal factors, such as time of day.
Figure 12 presents the average feature values strongly correlating with the "Not Tired" level across various locations. Data were collected exclusively from Haram, Mina, Arafah, and Muzdalifah, revealing distinct patterns in participants’ experiences. Participants recorded an average of 2,723 steps in Haram, 4,652 in Mina, 2,299 in Arafah, and just 139 in Muzdalifah. The average walking distances were 3,540 meters in Haram, 5,471 meters in Mina, 2,461 meters in Arafah, and 0 meters in Muzdalifah, as participants typically stayed in place during the overnight period.
The average step lengths varied slightly across locations: 0.54 meters in Haram, 0.59 meters in Mina, 0.56 meters in Arafah, and 0.54 meters in Muzdalifah. Additionally, average active energy expenditure demonstrated significant variation, with participants burning 183 kcal in Haram, 229 kcal in Mina, 82 kcal in Arafah, and only 19.25 kcal in Muzdalifah. These findings highlight the impact of location on physical activity and energy expenditure among participants.
Figure 13 illustrates the average features associated with the "Slightly Tired" level across locations, specifically focusing on data from Haram, Arafah, and Muzdalifah. This focus is deliberate, as Mina is predominantly associated with participants reporting "Extremely Tired" levels due to its physically demanding activities. Examining these other locations can give us more precise insights into the factors contributing to a "slightly Tired" status.
Participants averaged 7,193 steps in Haram, 4,946 in Arafah, and 3,944 in Muzdalifah, indicating higher activity levels in Haram, likely due to the nature of the rituals performed there. In contrast, activity decreased in Arafah and Muzdalifah. The average steps lengths varied slightly, recorded at 0.55 meters in Haram, 0.57 meters in Muzdalifah, and 0.59 meters in Arafah. The longer step length in Arafah may indicate more deliberate strides, possibly due to less crowding than in Haram.
In terms of walking running distance, participants covered an average of 8,193 meters in Haram, 5,722 meters in Arafah, and 4,828 meters in Muzdalifah. These figures reflect the different physical demands of each location, with Haram requiring more movement, while activities in Arafah and Muzdalifah were more focused on rest and prayer.
Participants burned an average of 387 kcal in Haram, 247 kcal in Arafah, and 87 kcal in Muzdalifah. This indicates that the physical demands were highest in Haram, where active engagement was more prevalent. In Muzdalifah, the lower energy expenditure suggests focusing on rest and less physically demanding activities, while Arafah reflects a moderate activity level.
Figure 14 illustrates the average value of features highly correlated with the ’Very Tired’ level across locations. Focusing exclusively on data from Haram, one of the most physically demanding locations, participants averaged 12,788 steps in Haram, with an average step length of 0.57 meters and an average walking running distance of 13,411 meters. They burned an average of 507 kcal, indicating significant physical exertion in this location.
Figure 15 presents the average values of features correlated with the "Extremely Tired" level across locations. Focusing on data recorded in Mina and Muzdalifah, as these locations are associated with some of the highest levels of physical exertion and emotional strain, participants averaged 12,644 steps in Mina and 14,661 in Muzdalifah, with average step lengths of 0.62 meters in Mina and 0.59 meters in Muzdalifah. Walking running distance averaged 14,484.6 meters in Mina and increased to 16,093 meters in Muzdalifah. The average active energy burned was 406 kcal in Mina and 594 kcal in Muzdalifah, reflecting the heightened physical exertion in these locations.
Figure 16 illustrates the relationship between stress levels and highly correlated features, accounting for temporal factors such as time of day and specific days during the pilgrimage. Understanding these relationships is essential for evaluating how pilgrimage activities impact stress levels. In Fig. 16, stress levels reach a "Good" rating instead of a "Very Good" rating when walking step lengths exceed 0.55 to 0.60 meters. The minimum step length corresponding to a ”Good” stress level was recorded on the 7th day of Hajj (Day 1), while the maximum step length was recorded on the 9th day, particularly later in the day. This pattern likely arises from participants using transportation to travel from Arafah to Muzdalifah, which may limit sustained walking and reduce stress levels, thus mitigating the potential negative impact of increased physical activity. A similar trend was noted on the 12th day (Day 5), where participants maintained a "Good" stress level with step lengths between 0.55 and 0.59 meters, particularly towards the end of the day. Overall, the data indicate that while longer step lengths are associated with a ”Good” stress rating, the unique circumstances of Hajj, characterized by intense physical exertion and intermittent transportation, significantly shape participants’ stress responses, underscoring the importance of considering environmental factors in this relationship.
[See PDF for image]
Fig. 17
Stress levels across each day in the pilgrimage
Figure 16, depicts the relationship between walking running distance and stress levels, revealing a clear transition from ”Very Good” to ”Good” when distances exceed 5,000 meters. This trend appears on the 7th, 9th, and 12th days of Hajj, with participants consistently reaching a ”Good” stress level by the end of each day. Generally, participants experience a ”Bad” stress level (the highest stress recorded during the pilgrimage) after covering 10,000 meters on all days, except on the 9th, when ”Bad” stress occurred after walking over 5,000 meters. This anomaly likely arises from the emotional strain typical of the 9th day, leading to heightened stress even at shorter distances. The emotional and psychological demands of the pilgrimage on this day are considerable, impacting stress levels despite lower physical exertion. These findings suggest that while longer distances generally correlate with higher stress levels, specific factors unique to the Hajj experience (such as fatigue and emotional strain) significantly influence participants’ stress responses.
Figure 16 examines the correlation between stress levels and the number of steps taken. Analysis reveals that stress levels increased to "Good" when participants exceeded 5,000 steps on key days such as the 7th, 9th, and 12th days of Hajj. Participants achieved step counts of up to 20,000 on these days, marking the peak for the "Good" stress level. The consistency of this pattern across multiple days suggests that increased physical activity correlates with stress management up to a certain threshold. Participants reported heightened stress ("Bad" level) after surpassing 10,000 steps on the 7th, 9th, and 10th days. Notably, at the end of the 9th day, participants experienced ”Bad” stress after reaching only 5,000 steps, suggesting that contextual factors, such as anticipated activities and cumulative fatigue, may overshadow the quantitative aspects of physical exertion. Factors such as emotional stressors, fatigue, and the nature of the pilgrimage activities likely contribute to the discrepancy between step count and stress levels. On the 10th day, participants recorded the highest step count associated with "Bad" stress (approximately 16,000 steps), highlighting the complex relationship between physical activity and stress.
Figure 16 shows how active energy expenditure correlates with stress levels, with a clear upward trend as stress increases. On the 7th day, energy expenditure readings for the ”Good” stress level began at approximately 300 kcal and peaked at 450 kcal. On the 9th day, participants burned between 200 and 300 kcal, while on the 12th day, readings hovered around 400 kcal. In contrast, "Bad" stress levels were observed on the 7th, 9th, and 10th days, with energy expenditures ranging from 450 to 600 kcal on the 7th day, dropping to just 100 kcal on the 9th, and fluctuating between 320 and 600 kcal on the 10th. These findings indicate that increased active energy expenditure is linked to heightened stress levels, highlighting that excessive exertion can exacerbate stress and negatively impact well-being. However, the lower energy expenditure on the 9th day (100 kcal) in high stress levels suggests that psychological factors or emotional stressors, may significantly influence stress outcomes.
[See PDF for image]
Fig. 18
Stress levels across each location in the pilgrimage
[See PDF for image]
Fig. 19
Stress levels across each stage in the pilgrimage journey
The analysis of stress levels throughout the pilgrimage reveals that participants reported the highest stress on the 7th, the end of the 9th, and the 10th days of Hajj, describing their feelings as ”Bad.” The 7th day exhibited a wide range of stress levels, while on the 9th day, they reported a mix of "Very Good" and "Good" stress levels and, at a certain point, "Bad" stress levels. In contrast, on the 11th day, most participants indicated an "Okay" stress level, as illustrated in Fig. 17.
As shown in Fig. 18, the analysis of stress levels at various pilgrimage locations reveals that Haram and Muzdalifah experienced the highest stress levels, with participants reporting a "Bad" level. In contrast, Arafah exhibited a predominantly "Good" stress level, while Mina recorded an "Okay" stress level.
The relationship between stress levels and pilgrimage stages is illustrated in Fig. 19. Participants experienced the highest stress levels during the Sa’i stage, with "Bad" being the most frequently recorded. Upon reaching Muzdalifah, they again reported the highest "Bad" stress level. In contrast, other stages, such as Tawaf and Standing in Arafah, exhibited a range of stress levels, with "Good" being the most commonly reported. The remaining stages mostly recorded a "Very Good" stress level, indicating an overall positive experience. This pattern suggests that the most physically demanding stages and crowded environments contribute to higher stress, while moments of reflection and prayer foster a more positive emotional state.
Figure 20 represents the correlation between the time of day, locations, and participants’ stress levels. By analyzing how stress fluctuates across different times and locations, we can determine whether certain times of day or locations contribute to heightened stress.
As shown in the figure, Arafah in the afternoon stands out, with approximately 25% of participants reporting a "Very Good" stress level, while the majority (around 75%) reported "Good". In Haram, stress levels remained consistent across the morning and afternoon, with most participants reporting "Very Good" in both periods. However, during the evening, around 15% of participants reported feeling "Good", while 5% felt "Bad". Muzdalifah exhibited the highest stress levels in the evening, with approximately 25% of participants reporting "Bad" stress levels, suggesting significant stress during this time. In contrast, 15% of participants in Mina reported an "Okay" stress level in the evening, while the majority experienced "Very Good" stress levels.
This analysis suggests that temporal factors impact stress levels less than environmental factors. However, environmental challenges in specific locations, such as crowds and heat, significantly affect participants’ stress.
[See PDF for image]
Fig. 20
Correlation between the time of day, locations, and stress levels
[See PDF for image]
Fig. 21
The common reasons cited for stress among participants based on stress levels
The reasons for stress levels vary among participants, offering valuable insights into the factors influencing their experiences. Figure 21 illustrates the common reasons cited by participants. Those who reported feeling "Very Good" frequently cited comfort as the primary reason. Many also highlighted the excellent organization, particularly the well-maintained pathways, significantly enhancing their overall experience. Similarly, participants who reported feeling "Good" mentioned heavy crowding as a concern. However, despite challenges such as hot weather and slight fatigue, these factors did not significantly impact their stress levels.
Participants who reported feeling "Okay" most commonly attributed their stress to long walking periods, contributing to moderate discomfort and stress. In contrast, those who reported feeling "Bad" cited more severe causes, including fatigue, extended walking, long bus waits, and hot weather. These factors significantly impacted their comfort, with many participants highlighting the challenges posed by the physical and environmental conditions they faced during the pilgrimage.
The analysis of stress indicators during Hajj and Umrah revealed significant patterns that contribute to understanding pilgrim stress dynamics. The analysis used mixed-effects regression models to study the relationships between stress indicators and contextual factors while considering the nested data structure that includes participant measurements. The analysis showed that step count strongly predicted stress levels because walking more during rituals led to higher stress levels. The analysis revealed that crowd density emerged as a significant predictor which demonstrates how environmental factors like navigating through crowded spaces affect pilgrim stress. The research provides strong evidence to identify major stress triggers which helps develop specific intervention methods to improve pilgrimage quality (Fig. 22).
[See PDF for image]
Fig. 22
The common reasons cited for stress among participants based on locations
Future work
The research outcomes from this study lead to multiple directions for future investigation which will enhance both stress understanding and management practices during Hajj and Umrah. The main research direction focuses on adding more biomarkers to create a detailed system for monitoring physiological stress reactions. Future research will include salivary cortisol testing as part of the study design to obtain hormonal stress data which will enhance the current biometric measurements. The research will study advanced machine learning methods to develop better stress level prediction and classification systems. The preliminary Support Vector Machine (SVM) model tests achieved 78% accuracy in stress level classification using the gathered data. The research will concentrate on improving these models and testing different algorithms to boost predictive accuracy and establish real-time stress tracking capabilities. The study requires a larger participant group consisting of diverse pilgrims to validate and apply the research findings. The 2025 research study will validate results by selecting more than 500 participants through stratified sampling methods which will increase the generalizability of the findings. The proposed research directions aim to develop an improved framework which provides actionable stress management solutions for pilgrimage environments.
Discussion
The research findings from this study generate important implications which affect both Hajj and Umrah pilgrimage management policies and practices. The research findings show that step count and crowd density play a crucial role in stress levels which requires specific interventions to address physical and environmental stressors. The practical implementation of dynamic crowd routing strategies should occur in Mina during peak hours to manage high-density areas. Real-time data collection of crowd density and stress indicators enables authorities to guide pilgrims toward less crowded routes which improves safety and comfort while reducing stress. The placement of hydration stations at 5,000-step intervals throughout the Hajj route would help reduce the negative impacts of physical strain and heat exposure that pilgrims commonly face. The availability of simple water stations and rest areas helps prevent dehydration and fatigue which leads to better health outcomes for pilgrims. A real-time alert system should be developed to notify pilgrims when their physiological levels exceed specific thresholds including an energy expenditure of 450 kcal/hour. The alert system would notify pilgrims to take rest or seek medical assistance which would help avoid health complications. The study-based policy recommendations establish practical and innovative solutions which support Vision 2030 goals by improving crowd management and healthcare services and pilgrim well-being. The study integrates technological insights with practical interventions to create a safer and more supportive pilgrimage experience.
Research limitations
By integrating physiological metrics (e.g., heart rate, step count, energy expenditure) with environmental factors (location, temperature) and subjective experiences (via validated questionnaires), this study achieves a triangulated approach to stress analysis. This methodology provides a holistic understanding of stress dynamics, elucidating interactions between physical exertion, environmental stressors, and emotional states—a feat rarely accomplished in prior pilgrimage research. The synthesis of real-time biometrics, geospatial tracking, and qualitative feedback transcends conventional single-metric approaches (e.g., surveys), establishing a novel framework for tech-driven health monitoring in mass gatherings. While larger sample sizes are essential for generalizability, the study’s multidimensional methodology offers critical insights into a high-stakes, understudied context, paving the way for scalable, context-specific interventions in future research.
Conclusion
This study successfully demonstrates the potential of employing advanced technologies—such as wearable devices, machine learning algorithms, and remote sensing systems—to measure and analyze stress during the Hajj and Umrah pilgrimages. By integrating physiological, environmental, and contextual data, the research offers comprehensive insights into the multifaceted nature of stress in pilgrimage settings. Notable findings reveal that stress levels peak during physically intensive rituals, particularly in Mina and Muzdalifah, driven by prolonged walking, extreme heat, and crowd density. Conversely, locations like Arafah recorded lower stress levels, attributed to a calmer, more reflective environment. The study also identifies critical stress thresholds, such as walking distances exceeding 10,000 meters and active energy expenditure surpassing 450 kcal, which strongly correlate with reported fatigue and stress. These findings align with the role of innovative technologies in enhancing the pilgrimage experience. The research introduces a novel dataset that supports the understanding of stress dynamics and serves as a foundation for future advancements in stress detection and management during Hajj and Large-scale gatherings.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Adam, E; Meiland, F; Frielink, N; Meinders, E; Smits, R; Embregts, P; Smaling, H. User requirements and perceptions of a sensor system for early stress detection in people with dementia and people with intellectual disability: Qualitative study. JMIR Form Res; 2024; 8, 52248. [DOI: https://dx.doi.org/10.2196/52248]
Alam S, Alam MAU (2023) Embrace: Explainable multitask burnout prediction for resident physicians using adaptive deep learning. medRxiv. https://doi.org/10.1101/2023.06.24.23291864, https://www.medrxiv.org/content/early/2023/06/29/2023.06.24.23291864.full.pdf
Alharthi N, Alshamarani M (2021) Analyzing the characteristics of pilgrims during their residence in Makkah using machine learning algorithms
Al-Shaery, AM; Aljassmi, H; Ahmed, SG; Farooqi, NS; Al-Hawsawi, AN; Moussa, M; Tridane, A; Alam, MD. Real-time pilgrims management using wearable physiological sensors, mobile technology and artificial intelligence. IEEE Access; 2022; 10, pp. 120891-120900.
Alshamrani, M. Iot and artificial intelligence implementations for remote healthcare monitoring systems: A survey. J King Saud University-Comput Inf Sci; 2022; 34,
Bae GE, Choi A, Beom JH, Kim MJ, Chung HS, Min IK, Chung SP, Kim JH (2021) Correlation between real-time heart rate and fatigue in chest compression providers during cardiopulmonary resuscitation: A simulation-based interventional study. Med 100(16):25425
Can, YS; Arnrich, B; Ersoy, C. Stress detection in daily life scenarios using smart phones and wearable sensors: A survey. J Biomed Inf; 2019; 92, 103139. [DOI: https://dx.doi.org/10.1016/j.jbi.2019.103139] Accessed 2025-01-23
Chan, VC; Beaudette, SM; Smale, KB; Beange, KH; Graham, RB. A subject-specific approach to detect fatigue-related changes in spine motion using wearable sensors. Sens; 2020; 20,
Dalmeida, KM; Masala, GL. HRV features as viable physiological markers for stress detection using wearable devices. Sens; 2021; 21,
Ehrhart M, Resch B, Havas C, Niederseer D (2022) A conditional GAN for generating time series data for stress detection in wearable physiological sensor data. 22(16). https://doi.org/10.3390/s22165969
Elgamal M, Alshamarani M (2019) An analytical study of the most important characteristics of the guests of Al-Rahman and their representation using modern technologies to mine large data to support
Fauzi, MA; Yang, B; Blobel, B. Comparative analysis between individual, centralized, and federated learning for smartwatch based stress detection. J Personal Med; 2022; 12,
Gao, Z; Xiao, X; Carlo, AD; Yin, J; Wang, Y; Huang, L; Tang, J; Chen, J. Advances in wearable strain sensors based on electrospun fibers. Adv Func Mater; 2023; 33,
Gedam S, Paul S (2021) A review on mental stress detection using wearable sensors and machine learning techniques. IEEE Access. 9, 84045–84066. Conference Name: IEEE Access. Accessed 2025-01-23. https://doi.org/10.1109/ACCESS.2021.3085502
Hosseini S, Gottumukkala R, Katragadda S, Bhupatiraju RT, Ashkar Z, Borst CW, Cochran K (2022) A multimodal sensor dataset for continuous stress detection of nurses in a hospital. 9(1):255. https://doi.org/10.1038/s41597-022-01361-y
Iqbal T, Elahi A, Wijns W, Shahzad A (2022) Exploring unsupervised machine learning classification methods for physiological stress detection. Front Med Technol 4. https://doi.org/10.3389/fmedt.2022.782756. (Publisher: Frontiers. Accessed 2025-01-23)
Kiranashree BK, Ambika V, Radhika AD (2021) Analysis on machine learning techniques for stress detection among employees. Asian J Comput Sci Technol 10(1):35–37. https://doi.org/10.51983/ajcst-2021.10.1.2698. Number: 1. Accessed 2025-01-23
López M-J, Arias CP, Romeu J, Jofre-Roca L (2023) Supervised machine learning-assisted driving stress monitoring mimo radar system. IEEE Sens J 23(23):28899–28911. https://doi.org/10.1109/JSEN.2023.3326880. (Conference Name: IEEE Sensors Journal. Accessed 2025-01-23)
Pereira NG, Dos Santos AM, Shinjo SK (2024) Association between wearable device use and quality of life in patients with idiopathic inflammatory myopathies and primary systemic vasculitis. Cureus 16(4)
Pote R, Jain S, Singatwar A, Yadav J, Vyas T, Kamdi A (2024) Emotional health analysis. International Research J Modern Eng Technol Sci 06(05). https://doi.org/10.56726/IRJMETS57170
Rashid, N; Mortlock, T; Faruque, MAA. Stress detection using context-aware sensor fusion from wearable devices. IEEE Internet Things J; 2023; 10,
Schmidt P, Reiss A, Duerichen R, Marberger C, Van Laerhoven K (2018) Introducing WESAD, a multimodal dataset for Wearable Stress and Affect Detection. In: Proceedings of the 20th ACM International Conference on Multimodal Interaction. ICMI ’18, pp 400–408. Association for Computing Machinery, New York, NY, USA. https://doi.org/10.1145/3242969.3242985. Accessed 2025-01-23
Shi, S; Cao, Z; Li, H; Du, C; Wu, Q; Li, Y. Recognition system of human fatigue state based on hip gait information in gait patterns. Electron; 2022; 11,
Talaat, FM; El-Balka, RM. Stress monitoring using wearable sensors: IoT techniques in medical field.; 2023; 35,
Yu H, Sano A (2022) Semi-supervised learning and data augmentation in wearable-based momentary stress detection in the wild. Submitted 22 February 2022. https://doi.org/10.48550/arxiv.2202.12935, https://arxiv.org/abs/2202.12935
© The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.