DEVELOPING AN INTEGRATED FRAMEWORK FOR RISK MANAGEMENT IN POLICING: CRISIS MANAGEMENT USING BIG DATA ANALYTICS AS A CASE STUDY

Objective: The study seeks to develop an integrated framework for enhancing risk management and crisis management in policing through the utilization of big data analytics, using the Dubai police force as a case study. Theoretical Framework: This research is grounded in the intersection of big data analytics and risk management within the field of policing, focusing on how data-driven approaches can innovate and improve security crisis management. Method: Data for the study was collected via questionnaires distributed to 450 police officers across all departments in Dubai. The analysis was conducted using AMOS software within a Structural Equation Modeling framework to assess the impact of big data analytics on police risk and crisis management. Results and Discussion: The results indicate that big data analytics significantly enhances risk management and crisis management in policing. It was found that police risk management serves as a mediating factor between big data analytics and security crisis management. These findings suggest that the application of big data can substantially improve knowledge and operational performance in police organizations. Research Implications: The study emphasizes the crucial role of big data analytics in revolutionizing police risk and crisis management processes. It highlights the potential for these technologies to provide substantial improvements in the efficiency and effectiveness of police operations. Originality/Value: This research is valuable as it provides empirical evidence of the benefits of big data analytics in the context of policing. It offers practical guidelines to police departments, particularly within the UAE, on leveraging big data to enhance their operational performance in managing risks and crises. The study contributes to the broader understanding of integrating advanced data analytics into public safety and emergency


INTRODUCTION
To date, there is no common agreement on the best way to deal with security crises.
Every country have their own view and perspective to deal with security crises.Security crises, e.g., intelligent leaks of confidential data that breach the national security, stealing data that affect the safety of people in the country, all these crises must be addressed and managed properly and swiftly.Given the location of the UAE in the Middle East, security studies and scholars in this field have investigated the succession of possible threats to national security through traditional lens, the have neglected the advancement in criminal methods that become very complicated and dependent on latest technologies, including the utilization of internet to hack confidential data that may harm the national security of UAE and become a source of scrutiny crises (Alzahrani, 2023;Guéraiche and Alexander, 2022).Therefore, facing crises and raising awareness among public entities like police departments is necessary to avoid further materials and losses (Kamil, 2020).Often, UAE national security crises policy is analyzed using standards from abroad, but it should be based on empirical evidence from local sources or through survey in specialist agencies (Salisbury, 2020).In addition, despite the origin of crises could be hard to identify to most people living in the UAE, the public view circumstances, not UAE police, as being responsible for continuing in any future crisis (Tariq, 2021).It is not clear to the present to what extent big data analytics and risk management contribute to security crises management.This gap has not been reported yet in the UAE.In brief, the potential of Dubai police in using big data analytics and risk management to control a security crisis has not been reported in the past, this scenario may give an ambiguous picture to the officials in UAE government when developing a guidelines and policy for security crisis management.
The primary significance of this study is filling the gap in the literature concerning the role of risk management and big data analytics in law enforcement organizations and understanding the direct and indirect effect of big data analytics in police departments (Alshamsi et al., 2024).In addition to that, this study is the first attempt to develop a crisis management framework in police departments.Hence, the result of this study is expected to improve the performance of the Dubai police departments.The development of risk management in police departments will enhance security in the country and provide better Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study 4 protection to the citizens against security threats.Moreover, the findings of this study should bring advantages to government agencies in UAE, especially those working in security fields and crisis management.This study studies the impact of Big data analytics, Police risk management on security crisis management: , a very recent topic where only few studyes were conducted earlier.

THEORETICAL LITERATURE AND HYPOTHESIS DEVELOPMENT
A literature review on the diffusion of innovations (DOI) theory indicates that it can serve as a theoretical framework for studying factors shaping utilization and the adoption of big data in developing countries.In crises, DOI proved to be a solid theory to explain how organizations use innovation to manage crises and risks (Alana and Sandra, 2021).This study deploys the DOI concept as a theoretical lens to offer a structure for the innovation process related to big data analytics and crisis management.DOI theory provides a framework for exploring how organisations practice and utilise innovations.Moreover, DOI explains the factors and processes influencing the adoption of innovations (e.g.big data analytics in law enforcement).

BIG DATA IN UAE POLICE
Big Data has been used in policing to improve the decision-making process in the daily operation of the police.A big-data-driven system that is sued to accurately dispatch the patrol cars of police in a geographic environment has been widely used in developing countries like UAE.For example, the big data systems are used to allocate, in real-time, the nearest patrol car to the location of an incident or crime scene.In recent years, this system has been implemented and applied widely by the Abu Dhabi police (Oualid, 2014).The UAE is leading regionally in investments and hosting cloud data centers that represent the future infrastructure of the electronic government, private and future technology sectors are based on big data at present, such as smart cities and self-driving cars, as well as the great benefits of these centers on the communications and information technology sectors in general, while police departments in UAE started to open new centers for big data analytics Al-Khaleej, 2021).In this regard, Abu Dhabi police are keen to continue in developing security and safety in the city, which confirms the critical role and approach in supporting Abu Dhabi to lead all cities in the world in the indicators of least crime cities for the fifth year in a row, through the continuous efforts made, to implement The principle of the rule of law, justice in all aspects of life, and tolerance among all residents, including citizens, residents and visitors (Al-Khaleej, 2021).
In the same context, Dubai Police has taken the initiative to develop an integrated system for examining electronic evidence using the feature of artificial intelligence and big data, which consists of a database of all electronic evidence issues that we deal with daily.Moreover, the police officers can rely on some advanced technical characteristics to speed up the processes used in analyzing data related to various crimes, stressing that Dubai police is currently investing in artificial intelligence and advanced technology based on big data and harnessing them to serve the reality of investigating forensic evidence and detecting crimes (Desk, 2016).
There are vital issues specifically related to the adoption of big data that must be addressed and acknowledged (Elyjoy, 2015).The rise of big data nowadays comes with many challenges, such as how organizations develop the necessary tools and methods for reacting to a crisis and misusing the vast amount of data for their advantage (Hassani and Silva, 2015).While new data technologies can develop police performance and law enforcement efficiency, concerns were raised regarding the way that police deal with big data (Alexander and Marion, 2020).
The Dubai Police Force is renowned for accepting new technologies to enhance law enforcement and protect the local people and the Dubai community (Desk, 2016).Today, Dubai police are investing in big data for crime prediction, the predictive policing software of Space Imaging Middle East (SIME), now has Dubai police force use on a large scale.The software analyzes existing intelligence and crime patterns from police databases and, using sophisticated algorithms and big data analytics, produces highly accurate data related to when and where crime is likely to occur next.The software, which is the first of its kind in the region, was developed in support of the UAE's Smart Governance Initiative, and specifically designed to complement the Dubai Police force's modernized approach to crime prevention and improved public safety.This intelligence informs patrol teams on which districts may require additional resources to prevent potential criminal activity.
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study Big data refers to both massive data sets and the instruments used to modify and analyze a vast volume of data.On the other hand, this term does not just refer to the information gathered from various sources; it also relates to the reasons behind the collected information.When data is gathered in bulk using algorithms to handle them (a set of instructions that tell a computer what to do), cross-reference data both inside and between datasets, the computational software that processes the data finds patterns.Big data is essential for many applications related to law enforcement, such as predictive policing by building discovering patterns about crimes (Daniel and Akwasi, 2019).Big Data has started to permeate all facets of our lives.With the increased interest in the topic of big data (Akter and Wamba, 2017).In addition to that, in the field of law enforcement, big data analytics holds the potential of uncovering previously unnoticed securityrelated patterns and revealing unanticipated hidden knowledge that could be the key to preventing future crimes or terrorist attacks (Aradau and Blanke, 2017).As a result, governments have made significant investments in big data collection systems and technology for acquiring and analyzing data (Hollin, 2015).These facts reveal how critical big data analytics are for policing missions and the great support the big data analytics can provide for police officers and all related police departments to predict crimes and manage security crises at the national level (Hand, 2009).Big data analytics is a multidimensional concept.Laney, (2001) proposed that the three dimensions of big data are volume, variety, and velocity.To describe big data, the three Vs have been utilized as a common framework (Gartner, 2015).In this section, the researchers will present the 3-Vs and other dimensions of big data provided by the computing industry as described below.

Data Volume
The amount of data collected and/or generated by an organization or an individual is referred to as volume.While 1 terabyte is presently the minimum size to qualify as big data, the minimum size to qualify as big data is a result of technological development.Currently, 1 terabyte contains enough data to fill 1,500 CDs or 220 DVDs, or roughly 16 million Facebook photos (Gartner, 2015).Unstructured data such as audio, pictures, and video is generated in large quantities by e-commerce, social media, and sensors.As more computing devices connect to the internet, new data is being added at an increasing rate.The rate at which data is generated and processed is referred to as velocity (Laney, 2001).
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study 7

Data Velocity
Data velocity rises over time.Due to the sluggish and expensive nature of data processing, firms initially evaluated data using batch processing systems.Real-time processing became the norm for computing applications as the speed of data production and processing grew.According to Gartner, (2015), there will be 6.4 billion connected devices in use worldwide in 2016, rising to 20.8 billion by 2020.Every day in 2016, it was predicted that 5.5 million new devices would be connected to collect, analyze, and share data.The increased data streaming capabilities of linked gadgets will further accelerate the velocity (Laney, 2001).

Data Variety
The amount of data kinds is referred to as variety.Organizations can now generate structured, semi-structured, and unstructured data thanks to technological advancements.
Unstructured data includes text, photos, audio, video, clickstream data, and sensor data, all of which lack the standardized framework essential for effective computation.Semi structured data does not comply to relational database specifications, but it can be designed to fit the structural needs of applications.Extensible Business Reporting Language (XBRL), which was developed to share financial data between enterprises and government agencies, is an example of semi-structured data.Structured data is predefined and can be found in a variety of standard database types.Unstructured data is generated at a far faster rate than structured data as new analytics tools are developed, and the data type becomes less of an impediment to analysis.
IBM added a fourth dimension, veracity, to highlight the unreliability and uncertainty inherent in data sources.Data incompleteness, inaccuracy, delay, inconsistency, subjectivity, and dishonesty cause uncertainty and unreliability (Laney, 2001).

POLICE RISK MANAGEMENT
Risks exist in every area of life and almost all industries, the business world, public administration, and law enforcement.Thus, decisions must be made constantly to avoid risks using all possible means and technologies (Roman, 2020).Understanding the uncertainty is an essential matter to ensure the suitability of the work.However, risk management as a discipline has been developed to help organizations identify and control risks through systematic approaches and practices.A risk in this context can be defined as a random event or incident 8 that may and may not occur, so if it does occur, it would have an undesirable impact on the organization's objectives or work (Vose, 2008).Different disciplines have different ways of classifying risks.Standard risk classification consists of three categories: 'known unknowns', 'known knowns', and 'unknowns'.These three categories correspond to different levels of uncertainty (Jorion, 2009).
Risk management is the process of identifying adverse events and estimating their likelihood of occurring with systematic preparation in advance.By running simulations and random variables with risk models, such as scenario tables, a risk manager can evaluate the probability of the best-and worst-case outcome, threat occurring in the future, and the damage the organization would experience should this threat become true (Adam, 2021).Like any other profession in policing, risk management requires striking a balance between achieving goals and avoiding the inevitable security threats.Protecting the public and battling crime entails operational risk, and exposing anyone to high levels of personal risk, whether or not they are police officers, can lead to stress-related anxiety.For many law enforcement agencies, risk management is a practice that seeks to identify and mitigate risk for both police officers and the public.It is essential to sustain the well-being of police officers and the public and ensure the integrity of the law enforcement institutions that protect and serve the community.Police officers engage in various risky activities every day, which involves several potential threatening situations.Some risks are categorized with the organizational level, such as financial (e.g. the high cost of crimes control), and physical, such as occupational health (Adrian, 2014).The ultimate impact of these risks is not just facing police officers but could affect the whole police department.While the effect of security risks is not limited to those associated with policing but also extends outward to the citizens and justice seekers.Thus, RM practices are essential in law enforcement organizations to protect the people and police officers (Adrian, 2014).Risk identification, risk assessment, and risk management methods should be viewed as an excellent, albeit restricted, approach to improving the likelihood of identifying and avoiding future offending or victimization (Alexander and Marion, 2020).A fully automated machine learning algorithms methods help police deals with risk and control crimes more effectively.However, the algorithms used in policing is only as good if the data used by these algorithms are accurate (Alexander and Marion, 2020).
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study 9

Risk Identification
Risks and uncertainties are two of the most widely used concepts in the project management literature (Adrian, 2014).Although these terms are closely related, many writers distinguish them (Alexander and Marion, 2020).It is also difficult for workers at risk to identify and distinguish them.The definition of risk or uncertainty regarding the use of a particular project is often adjusted.Roman (2020) suggested that identification of risk is critical.He added that uncertainty is the intangible measure of what we don't know.Uncertainty is left behind when all the potential risks have been identified.Uncertainty is gaps in our knowledge we may not even be aware of.Therefore, risk must be determined before establishing any projects.

Risk Analysis
The second stage of the RM process is risk analysis, which involves analyzing the data obtained concerning the potential risk.Risk analysis may also be defined as the process of selecting the chances that have the most significant influence on the work from all of the threats identified during the identification phase (Alexander and Marion, 2020).
In policing, risk analysis affords support for risk practices about how factors combine to increase the likelihood of crime occurrence (Kennedy, 2018).Conventional risk analysis methods tend to underestimate the probability and impact of risks (e.g.pandemics, financial collapses, terrorist attacks), as sometimes the existence of independent clarifications is wrongly assumed and cascading errors that can occur in complex systems are not considered (Roman, 2020).Risk analysis help organizations to forecast the future with confidence; it is fundamental to predict uncertainties and reduce their incidence or impact.The adoption of risk management increases the likelihood of successful work completion and decreases its risks.

Risk Evaluation
In general, there are two approaches of risk evaluation in the literature: (1) quantitative (2) qualitative.The first approach is based on data, whereas the second approach is based on interviews and brainstorming techniques (Beieler et al., 2016).It is vital that the most probable risk factors are effectively quantified in advance and next analysed before using a qualitative approach.Exposure to potential risks can result in delays, reduced productivity, and increased operating costs for many others.Many institutions are already adopting innovative ways to use 10 big data analytical methods to improve their risk assessment processes and predict economic, social, or environmental data (Beieler et al., 2016).Risk evaluation is associated with practicing insight into the dangers an organization or staff faces in a specific location.A risk evaluation in the mission of law enforcement is a fundamental element and must be viewed as an integral part of the broader assessments involved in establishing operations or programs in any police department (Bickley, 2017).Evaluating the risk must not be a one-off event.A continuous reevaluation of all possible risks will help ensure that you have appropriate security measures in place at all times.The risk assessment process first identifies the different security threats within a given context and how your staff, assets, the programs being implemented, or the organization could be vulnerable.

Risk Response
This last stage in the process of RM indicates the actions to respond to the risk.The strategy and approach chosen for risk response depend on risk types (Vose, 2008).Risk response includes specific information and more details about the path taken to respond to the risk (Kennedy, 2018).The main requirement for risk response is that the risk manager must have adequate knowledge to respond to the risk (Kennedy, 2018.Some risk strategies for risk response include avoidance, reduction, transfer and possession (Bickley, 2017).In addition to these types of reactions, Vose, (2008) also describes how fast risk response should be in the typical case of RM.He suggested responding to the risk effectively, and it is essential to collect relevant information to prevent the risks.The literature review in this section indicates that RM is a multi-dimensional variable.The findings from previous studies reveal that RM can be measured through several dimensions.It is found that some dimensions are widely cited in the literature, such as risk identification, risk analysis, risk evaluation, and risk response.In contrast, few studies identified other dimensions not well explained or quantitatively measured in study associated with crisis management.
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study Security and crisis management establish a technique, system, and structure to protect people, property, and other assets from harm, loss, or criminal activity.This guarantees that police officers better grasp potential security risks and know how to respond to threats and safety emergencies.
According to Khaddam (2014) claimed that crisis management has three principal stages: pre-crisis management stage, crisis management stage (during the crisis), and post-crisis management stage (after the crisis).In the same context, Bundy et al., (2017) provided a model that included five stages of crisis management, what the organization must do at each step, and the duties and tasks that must be worked out to be appropriate for each stage.As a result, the organization now has the required information to do this assignment successfully.These steps will be regarded as dimensions of crisis management for this study: Detection stage: This refers to the stage before the start of the crisis, and it is represented by the organization's ability to respond to the warning signals that may cause the crisis, which includes taking preventive measures to prevent the crisis from occurring, or at the very least reducing the severity and effects of the crisis if it does happen despite the organization's efforts to avoid it, and discovering early warnings of crisis occurrence (Khaddam, 2014).
Prevention Stage: At this stage, organizations must develop an emergency plan to address the crisis and assemble a crisis management team with as much prior crisis management expertise as possible.This stage also entails identifying the organization's weak points and devising a strategy to address them, as well as identifying the necessary methods and tools to assist in dealing with the crisis, training employees on how to deal with crises, learning from the mistakes of others, developing information about crises and expected problems, and assessing their skills in crisis management (Khaddam, 2014).
Containment stage.At this stage, a set of actions to be taken is defined, communication processes within the crisis field are organized, the situation is stabilized, losses are reduced, the psychological and social effects of the crisis are addressed, and functional performance is improved in a more effective way than before.As a result, the plans are carried out by enforcing the order to deal with the problem, reducing the crisis, and finally utilizing the organization's resources (Saffar and Obeidat, 2020).
Recovery Stage.At this stage, the organization's activity and operations are resumed.
The organization evaluates its loss and loss and assesses what is required to benefit from the activity and balance entirely.Human resources play an essential role in this process, as the organization's human resources situation must be evaluated.What are the remaining capabilities of the organization, and what are the assets used and exploited after knowing the resources (Saffar and Obeidat, 2020).
Learning Stage: Also known as the stage of drawing morals and lessons from prior crises to construct experiences capable of avoiding crises, preventing their recurrence, and standing at flaws, improving and avoiding them through the development and improvement process (Bundy et al., 2017).
The nature and tasks of the crisis management team vary depending on the type of incidents/situations, their location, and the level of assistance necessary.One crisis management team member may perform multiple functions (Laari-Salmela et al., 2019).
Additional support responsibilities, such as security, finance, insurance, legal counsel, social media, internal communications, and IT, may be necessary depending on the nature of the crisis and organizational capabilities.
The literature review in this section indicates that crisis management is a multidimensional variable.The findings from previous studies reveal that crisis management can be measured through several dimensions.It is found that some dimensions are widely cited in the literature, such as crisis detection, crisis prevention, crisis containment, crisis recovery, crisis learning.In contrast, few studies identified other dimensions not well explained or quantitatively measured in study associated with crisis management.

The Relationship Between Big Data Analytics and Security Crisis Management
Crisis management systems are one of the big data roles, and machine learning can be used (Alpaydin, 2009).Scientists and analysts face one of the biggest challenges of managing large volumes of data generated during disasters and crises.Therefore, the role of big data in crisis management has been evolved (Abderrazak et al., 2020).The literature revealed that big data analytics directly impacts crises management in various fields and industries.In particular, massive unstructured text big data, such as articles, blog posts, social media posts and comments, speeches, presidential inauguration, and web platforms' documents through search engines, can be easily analyzed with the development of analytical tools (Park, 2021).
According to the theory of crisis management, big data can lead the direction of potential crisis 13 management and has created opportunities for it to improve and control through the analysis of crisis information.There is great confidence that big data tools can be used to process large amounts of crisis-related data to provide an insight into the fast-changing situation and help drive an effective disaster response (Qadir et al., 2016).Furthermore, it is necessary to prepare for a crisis management strategy at the macro-level in each country to respond to crises that bring about social change.In this regard, many scholars are focusing on using big data to read trends and agendas to discover insights into big data using various data mining techniques for responding to the unknown crisis and future events (Yun and Park, 2018;Abderrazak et al., 2020).Park ( 2021) study shows how a country can use big data analytics to detect "key social issues" and then make a subsequent strategy or decision-making system to develop a public communication or new policy in future events.Hence, it is necessary to prepare for a crisis management plan within a short time in a crisis that brings about social change (Qadir et al., 2016).Watson et al. (2017) presented findings from a case study of big data roadmap and supports results from other studies in that big data can contribute to crisis response efforts.They concluded that the increased usage of large datasets could positively affect preparation and response to crises and disasters, massive data analytics.
In the same context, Watson et al (2017) suggested that big data analytics is incorporated into the knowledge management process to improve the capability of data processing and crises.
Their case study shows that the proposed knowledge management solution helps improve situational awareness and decision making when dealing with social security incidents.At the same time, Doka et al. (2017) found a strong association between big data applications and crisis control, especially crises associated with the riot in the community.Based on the empirical results and findings from previous studies.This study assumes that big data analytics and security crisis management are linked.This association suggests that big data analytics directly affects security crisis management.Build upon the above explanation.The hypothesis is formulated as follows: Hypothesis 1: "Big data analytics has a significant effect on security crisis management."

The Relationship Between Big Data Analytics and Police Risk Management
Some scholars suggested that big data and predictive analytics cannot ensure that all critical problems would be avoided before they occur, but big data analytics can provide more The amount of data acquired through digital technologies and multi-channelling with the adoption of big data analytics could support the maximization of global business value thanks to the alignment of strategic priorities for risk management activities, the timely reporting of sources of uncertainty on which to focus attention, and the implementation of specific actions to improve performance (Nyman et al., 2018).
Additionally, the complexity of the police mission poses a considerable challenge for risk analysis and forecasting.Conventional risk analysis methods in law enforcement tend to underestimate the probability and impact of risks (e.g.pandemics, terrorist attacks), as sometimes the existence of independent observations is wrongly assumed and cascading errors that can occur in complex systems are not considered (Roman, 2020).Whereas big data analytics offer substantial opportunities for improving risk management but may not replace the significance of appropriate assumptions, adequate data quality and continuous validation (Nyman et al., 2018).Although there are different understandings as to whether or not the main methods of risk management for large amounts of data are similar to conventional methods, it is widely considered that the availability of big data analytics allows novel risk management (Roman, 2020).This study will validate the following hypothesis statement based on the previous findings.
Hypothesis 2 : "Big data analytics has a significant effect on police risk management."

Management
Crisis management is not necessarily the same thing as risk management.Risk management involves planning for events that might occur in the future, while crisis management involves reacting to adverse events during and after they have occurred (Adam, 2021).As explained earlier, risk refers to the probabilistic likelihood that a crisis may happen and its (often economic) impact.Therefore, the risk is always linked to crises (Zamoum and Gorpe, 2018).
The increasing number of crises of all kinds requires preventive measures.One of the elements is risk management in crisis management (Skomra, 2017).Effective risk management can prevent an issue from becoming a crisis.Poor understanding and management of risks can lead to a crisis (COMCEC, 2017).The significant effect of crises is well known even on the security sector, especially the police (Tariq, 2021).Hence, to foster crises management in policing, risk management is one of the main requirements to achieve this goal.In crisis management, risk management has been conceived mainly as a proactive pre-crisis management effort deployed for crisis prevention and preparedness efforts (Coombs, 2015).
Thus organizations should continually identify, manage and communicate risks to key stakeholders during the different phases of crisis management (Ndlela, 2019).
In addition to crisis preparedness, risk assessment can inform other phases of the risk management cycle, including vulnerability reduction through long-term territorial management, infrastructures and other policies, as well as disaster risk financing strategies.It can constitute a fundamental tool to harmonize risk management policies and practices across its various components with an overall coherent vision of priorities (Charles, 2013).
In brief, crises are often characterized by uncertain elements and create new risks for organizations involved.Uncertainty surrounding the crisis circumstances poses many risks and significantly intensifies risk variables anchored around the probability of the event causing harm and the consequence of that harm (Zamoum and Gorpe, 2018).In other words, the crisis necessitates the identification of possible risks (including all types of threats and stakeholders associated with the crisis).After identifying and analyzing the risk issue, the organization has to decide how it intends to frame the issue (Ndlela, 2019).The previous findings show that the adoption of crisis management alone is not enough without adequate risk management.Still, this relationship has not been examined in the law enforcement domain/ Therefore, this study will investigate the following hypothesis to understand the connection between risk management and security crisis management in policing missions.
Hypothesis 3: "Police risk management has a significant effect on security crisis management."

The Mediation Role of Risk Management
The literature review shows that risk management, crisis management, and big data analytics are linked.However, how risk management interacts with big data analytics and security crisis management in policing missions arises.Risk management is critical for any organization, and in the big data era, analytical tools for risk management are evolving faster than ever (Ozgur et al., 2020).Facts showed that risk management and big data analytics could identify new risks from data patterns for effective risk management strategies and better crisis management in departments responsible for security decision-making.One of the main benefits of big data analytics is having a better risk management strategy that draws from large data sample sizes (Bakdash and Marusich, 2015;Buchanan, 2019).Although police may be rich in data, police still need to improve the extraction of information and knowledge from that data and use it to decrease crime and strengthen clearance rates (Ridgeway, 2018).While risk analysis provides evidence-based support for risk narratives about how factors combine to increase the probability of crime occurrence (Kennedy et al., 2108).Thereby, risk management influences the relationship between big data analytics and crisis management.Nevertheless, the mediation role of risk management on this association has not been examined empirically in the context of law enforcement.Based on this claim, this study will verify the following hypothesis statement.
Hypothesis 4: "Police risk management mediates the relationship between big data analytics and security crisis management."

FRAMEWORK OF THE STUDY
This study deploys the DOI concept as a theoretical lens to offer a structure for the innovation process related to big data analytics and crisis management.DOI explain how an organization uses innovation to communicate within the organization and the public.For example, police departments use innovative tools such as big data analytics to serve the public in specific crises and big events.The DOI Theory helps provide an account of how technological innovations such as big data move from the stage of the invention to widespread use or not (Elyjoy, 2015).To develop a framework to identify the main factors affecting the organizational adoption of big data, DOI is a robust theory to enhance practitioners' understanding of the decision-making processes involved in a firm's adoption of big data (Shiwei et al., 2018).A literature review on the DOI Theory indicates that it can serve as a theoretical framework for studying factors shaping utilisation and the adoption of big data in developing countries such as the UAE.DOI is very comprehensive, and its concepts are very relevant to technology adoption in developing countries (Roman, 2004;Aleke et al., 2011;Richardson, 2009).Therefore, the conceptual framework of this study is designed to explain how the innovation diffusion through big data analytics in law enforcement organizations could foster security crisis management with the mediation influence of police crisis management.
The framework that is developed in this study constructs one independent variable (big data analytics), one mediator (police risk management), and one dependent variable (security crisis management).The conceptual framework is a body of interconnected fundamentals and objectives.This framework will be evaluated based on quantitative methods and statistical analyses.Hence, this study aims to identify the impact of big data analytics on security crisis management and the underlying police risk management that achieve the objectives of the police mission in UAE.Those concepts will guide the Dubai police headquarter on how diffusion of innovation in the big data domain can support security crisis management in the future.

Figure 1
Study Framework

DATA AND METHODOLOGY
During the survey phase, the researcher has distributed questionnaires to a sample of police officers consists of 450 police officers in all police departments in the city of Dubai.Out of the 450 questionnaires distributed to the study sample, 394 valid questionnaires have been considered for the analysis.This implies a response rate of 87.55%.If a response rate less than 50% of the defined sample then lead to incorrect generalization of results with respect to the whole population.The remaining number (56 forms) represent the questionnaires that include missing data.Missing data is one of the most persistent problems during data analysis procedure.Moreover, the missing data could deviate the conclusions and should not be included in any analysis.Therefore, 56 incomplete questionnaires, such as missing data or blank are filtered and then removed from data analysis.Nonetheless, the minimum sample requirement for generalisation was met.A period of six (6) weeks was allocated to collect data.The results were downloaded from the online data collection platform and analyzed with Analysis of (AMOS).The collected data was originally entered into the online data collection platform based on the weights of the various responses.

DEMOGRAPHIC ANALYSIS
Table 1 shows the respondents profile for those participated in the research.
Demographic analysis includes the statistics that explain the characteristics of study sample.
Hence, this section demonstrate the population dynamics by investigating the main demographic data such as age, gender, academic level, work experience of the police officers as shown in Table 1.With respect to gender, the number of males is higher than females in the  The police officers reveal that big data analytics take a high priority in policing at the present day.The rise of big data in policing exposes the cutting-edge technology transforming how cops perform their jobs and demonstrates why it is more vital than ever for citizens to comprehend the far-reaching implications of big data analytics as a potent law enforcement weapon (Andrew, 2017).However, when police officers use big data and predictive analytics, their policing performance becomes more proactive, even highly responsive.In this regard, Daniel and Akwasi (2019) claimed that big data analytics could shorten the time it takes for police officers to arrest half by relying on cross-referencing databases compared to regular arrest routines.Similar rating has been obtained with respect to risk management and crisis management.Likewise, Roman (2020) found that police officers' safety and wellness are high priorities for managing risk in law enforcement for many reasons, not the least of which is that healthy officers are less likely to engage in risk-inducing behaviors such as excessive use of force and abuse of power.To that end, risk reduction strategies inspire police officers to experiment with more place-based involvements to overcome the complaints about personfocused policing and excessively aggressive (Kennedy et al., 2018).
In addition, the respondents reveal some concerns with respect to the obstacles that become problematic for crises management in Dubai police, especially the absence of clear strategy and Shortage of expertise in crisis management.This finding indicate why police agencies should use the experience of the necessary people from diverse operational areas in crisis management to plan and manage a security crisis as it happens.Experts who have the sufficient skills, constructive thinking, feeling energized and engaged with staff, knowing, behaving responsibly, and having confidence in the crisis team can all be precious qualities in good crisis management.Furthermore, understanding and applying these quantitative abilities allows crisis leaders to think creatively about difficulties, providing them with a better knowledge of the security situation (Khaddam, 2014).

CONSTRUCT VALIDITY ANALYSIS
Construct validity is essential to the perceived overall validity of the empirical data in this study.Convergent validity, along with discriminant validity, is a subtype of construct validity.It is a primary analysis that should be conducted before CFA.In this analysis the associations of the indicators are evaluated in relation to their related dimensions of big data analytics, risk management, and crisis management.The first step is exploring the true structure of indicators with their related dimension by evaluating the initial measurement models based on the dataset.In other sense, these steps are necessary to improve the fit of data with the measurement model before building the structure model.The convergent validity is associated with the indicators belong to a single dimension and reveals how close those indicators to determine a particular dimension.To establish convergent validity, the Average Variance Extracted (AVE) should exceed (0.50).Referring the output data in Table 2, it is evident that AVEs for all dimension within the standard threshold (≥ 0.50), with a condition that the composite reliability ≥ 0.7.In conclusion, each dimension is adequate to convey sufficient variance for indicators to converge into a single dimension.

Table 2
The amount of AVE and composite reliability of dimensions Next step is measuring the discriminant validity of constructs in order to know how far the dimensions that are related to a single construct distant from each other.In other words, the discriminant validity refers reveal the extent of differences between the dimensions of a single construct, as well as the level of non-correlation between them.The condition is that the AVEs related to each dimension must be greater than all the correlation between the dimensions of a single construct and less than 0.70.As shown in Table 3, all correlations are less then (0.70) the threshold point which ensure that discriminant validity of constructs has been obtained.

TRUCTURAL EQUATION MODELING
Reading the output data in Table 4. shows that the magnitude of CMN/DF range in between 1.474 -1.813.To achieve a good fitness with the observed data, the Normed-ratio (CMIN/D) is preferred to be less than 3.00.CFI magnitude ranges in between 0.966 -0.993 (cut-off ≥ 0.90).TLI magnitude ranges in between 0.965 -0.991 (cut-off ≥ 0.90).While PCLOSE is non-significant for all measures (Non-Sig.≥ 0.05), and RMSEA ≤ 0.08 for all constructs which indicates a good data fit with the measurement model.In conclusion, all fitindices are matching the cut-off points of SEM standards.

24
The second step in CFA analysis is evaluating the strength of relationships (i.e., beta) between the dimensions and their associated construct, as well as ensuring that all these relationships are significant as shown in Table 5.In conclusion, the CFA analysis reveal that all constructs are second order (consisting of a multiple sub-constructs or dimensions).Each sub-construct (dimension) is measured using certain number of indicators.Whereas big data analytics is linked to three latent dimensions (data volume, data variety, data velocity), Risk management is linked to four latent dimensions (risk identification, risk analysis, risk evaluation, risk response)..

HYPOTHESES TESTING
In Figure 2   This section presents the result of hypotheses testing for direct effect.The direct effect test aims to examine the relationships between the independent variables and the dependent variable.The results are presented in table 4 and explained in the following conclusions.As reported in Table 4.10; Hypothesis (H1) is accepted states that "Big data analytics has a significant effect on security crisis management" (Sig.= 0.00, C.R = 3.13 ≥ 1.96).In line with this findings, previous studies reported similar results.Big data analyticscan be used to process enormous amounts of crisis-related data in order to provide insight into the rapidly changing situation and help drive an efficient catastrophe response (Qadir et al., 2016).To that end, this result shows why police departments in each country must establish a macro-crisis management strategy based on big data in order to respond to security crises that cause social upheaval.Many scholars are focusing on using big data to read trends and agendas in order to discover insights into big data using various data mining techniques in order to respond to unknown crises and future events (Park and Alenezi, 2018).Likewise, Doka et al. (2017) found a strong association between big data analytics and crisis management, especially crises associated with the riot in the community which requires the interfere from the police.
In addition, Hypothesis (H2) is accepted as well, which states that "Big data analytics has a significant effect on risk management" (Sig.= 0.00, C.R = 4.61 ≥ 1.96).In the same context, several studies have indicated an important connection between risk management and big data analytics.The ability to communicate a significant volume and variety of information (i.e.Big Data) within an organization is critical for the interactive and multidirectional risk assessment and management process.The risk management process is critical in any work that involves risks.As a result, it is critical for every business that deals with risk on a daily basis and has a long-term objective to identify the competencies required of the risk manager in the digital age.Big data has significant prospects for enhancing risk forecasting, but it should not be used in place of acceptable assumptions, adequate data quality, and constant data validation (Hassani and Silva, 2015).In fact, risk analysis that is based on big data should help police to predict risks that affect the work of police and law enforcement, as well as the national security, for example, cloud computing, social media, Internet-of-things, quick data retrieval and storage, and so on.Today, police agencies in developed countries like US and Europe have access to unprecedented volumes and versions of data.As a result, risk analysis approaches are also advancing and breaking new ground.For example, enormous data on financial actions of corporations can help individuals, regulators, industries, and others make risky decisions (Choi and Lambert, 2017).The ongoing tracking of discrimination risk is needed at all stages of a police data analytics project, from problem formulation and tool design to testing and Finally, Hypothesis (3) has been accepted which states that "Risk management has a significant effect on security crisis management" (Sig.= 0.00, C.R = 4.21 ≥ 1.96).The review literature reveal similar findings.Risk management can constitute a fundamental tool to harmonize crises management (Baubion, 2013).In other words, crises are often characterized by uncertain elements and create new risks for organizations involved.Uncertainty surrounding the crisis circumstances poses many risks and significantly intensifies risk variables anchored around the probability of the event causing harm and the consequence of that harm (Zamoum and Gorpe, 2018).After identifying and analyzing the risk issue, the police has to decide how it intends to use risks management in favor crisis management (Ndlela, 2019).

Hypothesis Testing (Mediation Effect)
In any mediation model, the relationship between the independent and dependent variable is influenced by a third variable called the mediator which is "risk management" in the case of this study.As a result, when "risk management" indirectly influence the direct relationship between big data analytics and crisis management.The strength of direct relationship should be reduced by risk management (the mediator), but with a condition that this direct effect remains significant.As stated above, Table 5 which report the significance of direct relationships between big data analytics, risk management, crisis management.If one of these relationships is non-significant, then a mediation effect does not exist.Otherwise proceeding for testing the significance of indirect and total effect using bootstrapping technique.
The next step in mediation analysis is testing the indirect and total effects between the constructs so that to confirm the mediation role of risk management.Bootstrapping is a popular and reliable approach in empirical studies for testing mediation effects (MacKinnon et al., 2004).AMOS software is utilized for conducting bootstrapping.The "total effect" = "indirect effect" + "direct effect".It is necessary that all indirect and total effects are statistically significant in order to consider the mediation role of risk management.MacKinnon et al (2004) indicated that when the indirect impact of big data analytics on crisis management through risk management, with 95% bootstrapping confidence interval does not overlap a 0 in between, this indicates a mediation effect.It is found that the indirect (mediated) effect of big data analytics on crisis management is significant (Sig.= 0.03), and 0 out of interval (0.139, 0.832).These two conditions are essential to consider the indirect path big data analytics → crisis management is partially mediated by risk management.The same assumption is applied to assess the total effect between the independent and dependent constructs.In brief, the result from bootstrapping reveal that both effects (indirect/total) due to the partial mediating influence of risk management are statistically significant (Sig.≤ 0.05).Furthermore, the evaluation of output data in Table 7 shows that the (0-null) bound falls out of confidence intervals for all paths.Based on this result, hypotheses 4 which states that "risk management mediate the relationship between big data analytics and crisis management" is accepted and supported.

RESULTS AND DISCUSSION
The police officers reveal that big data analytics take a high priority in policing at the present day.The rise of big data in policing exposes the cutting-edge technology transforming how cops perform their jobs and demonstrates why it is more vital than ever for citizens to comprehend the far-reaching implications of big data analytics as a potent law enforcement weapon (Andrew, 2017).However, when police officers use big data and predictive analytics, their policing performance becomes more proactive, even highly responsive.In this regard, Daniel and Akwasi (2019) claimed that big data analytics could shorten the time it takes for police officers to arrest half by relying on cross-referencing databases compared to regular arrest routines.Similar rating has been obtained with respect to risk management and crisis management.Likewise, Roman (2020) found that police officers' safety and wellness are high priorities for managing risk in law enforcement for many reasons, not the least of which is that healthy officers are less likely to engage in risk-inducing behaviors such as excessive use of force and abuse of power.To that end, risk reduction strategies inspire police officers to experiment with more place-based involvements to overcome the complaints about personfocused policing and excessively aggressive (Kennedy et al., 2018).
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study behaving responsibly, and having confidence in the crisis team can all be precious qualities in good crisis management.Furthermore, understanding and applying these quantitative abilities allows crisis leaders to think creatively about difficulties, providing them with a better knowledge of the security situation (Khaddam, 2014).Moreover, big data analytics as a study discipline are still evolving and not yet established.Thus, understanding the phenomenon and the variables that interact with big data and crisis management still needs more investigation and in-depth quantitative analysis.At the same time, a framework that correlates big data analytics, risk management, and crisis management are yet to be established.The actual progress made in big data analytics revealed a lack of management study in the field and a distinct lack of theoretical constructs and academic consistency perhaps a function of an underlying methodological rather than an educational challenge.In other words, there has also been a lack of study studies that broadly address the critical challenges of big data in the UAE or investigate opportunities for new theoretical models (George et al., 2014).Thus, studies in this field should identify the big data challenges and associate big data analytics methods with crisis management to understand how these two variables are associated (Uthayasankar et al., 2017).
It is important to note that the availability of big data alone does not stop crisis or provide complete protection from the consequences of crisis (Bacon., 2013).A good example is the existence of a vast amount of data on earthquakes, but the lack of a consistent model that can precisely predict earthquakes (Silver, 2012).Some existing studies found that the challenges of crisis management are related to a hypothesis, testing and frameworks utilized for big data predicting (Poynter, 2013), whilst West (2013) recognized more concern on the absence of theory to complement big data in a wide range of industries.Besides, the diverse challenges linked to predicting big data should be given more consideration (Hassani and Silva, 2015).
Based on the previous claims, this study will examine the role of risk management on the direct relationship between big data analytics and crisis management for the first time in law enforcement agencies in the UAE.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH
Moreover, big data analytics as a study discipline are still evolving and not yet established.Thus, understanding the phenomenon and the variables that interact with big data and crisis management still needs more investigation and in-depth quantitative analysis.At the same time, a framework that correlates big data analytics, risk management, and crisis management are yet to be established.The actual progress made in big data analytics revealed a lack of management study in the field and a distinct lack of theoretical constructs and academic consistencyperhaps a function of an underlying methodological rather than an educational challenge.In other words, there has also been a lack of study studies that broadly address the critical challenges of big data in the UAE or investigate opportunities for new theoretical models (George et al., 2014).Thus, studies in this field should identify the big data challenges and associate big data analytics methods with crisis management to understand how these two variables are associated (Uthayasankar et al., 2017).It is important to note that the availability of big data alone does not stop crisis or provide complete protection from the consequences of crisis (Bacon., 2013).A good example is the existence of a vast amount of data on earthquakes, but the lack of a consistent model that can precisely predict earthquakes (Silver, 2012).Some existing studies found that the challenges of crisis management are related to a hypothesis, testing and frameworks utilized for big data predicting (Poynter, 2013), whilst West (2013) recognized more concern on the absence of theory to complement big data in a wide range of industries.Besides, the diverse challenges linked to predicting big data should be given more consideration (Hassani and Silva, 2015).
Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.2 | p.1-36 | e06035 | 2024.14 accurate early-warning indicators to prevent and reduce risks(Dahleh et al., 2016;Nyman et al., 2018).Hence, the adoption of big data analytics in risk management can create an essential competitive advantage for organizations.However, the direction of a highly variable amount of data in real-time requires new tools and methods and the broadening of IT, statistical and mathematical knowledge, mainly oriented to quantitative data analysis to interpret and transform it into high added-value information.The ongoing tracking of discrimination risk is needed at all stages of a police data analytics project, from problem formulation and tool design to testing and operational deployment (Alexander and Marion, 2020).
Dubai police, while those aged 40-49 years are the highest number of officers working in Dubai police, while those aged 60 years an above are the lowest number in population.The data of academic qualification reveals that officers having a master degree are the highest number, and those holding only PHD are the smallest group in Dubai police.In addition, officers who have middle experience (10-15 years) in law enforcement represent the highest percentage in Dubai police, whereas fresh officers (1-5 years) are the lowest number in Dubai police.
report that all fit-indices are compatible with the cut-off points for SEM standards.Starting with PCLOSE = 0.990 (perfect non-significant), while RMSEA = 0.046 (≤0.08) which reflects a high degree of model-fit.Moreover, CMIN/DF = 1.820 (≤ 3.00), CFI = 0.919 (≥ 0.80), TLI = 0.914 (≥ 0.80), the typical range for TLI and CFI lies between zero and one, whereas TLI and CFI values close to 1 indicate a very good model fit.These magnitudes are typical and shows that DandM model in HIS domain is valid.These are the standard fitindices in SEM standards that should be used to assess the validity of a theoretical model with an empirical data.
Figure 2Study Framework Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.2 | p.1-36 | e06035 | 2024.27 operational deployment (Alexander and Marion, 2020).These evidences should be used by Dubai police to enhance risk management based on big data analytics.
___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.2 | p.1-36 | e06035 | 2024.29 In addition, the respondents reveal some concerns with respect to the obstacles that become problematic for crises management in Dubai police, especially the absence of clear strategy and Shortage of expertise in crisis management.This finding indicate why police agencies should use the experience of the necessary people from diverse operational areas in crisis management to plan and manage a security crisis as it happens.Experts who have the sufficient skills, constructive thinking, feeling energized and engaged with staff, knowing,

Table 3
Correlations between Constructs

Table 4
Fit indices of measurement models Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.2 | p.1-36 | e06035 | 2024.

Table 5
Significance and strength of relationships between constructs and dimensions

Table 6
Standardized regression weightsTo accept the hypotheses, the Critical Ration (C.R) is applied to evaluate the significance level of unstandardized regression coefficients.The following two criteria are used to validate the hypotheses: (1) if C.R ≥ 1.96 for a specific relationship, then a hypothesis is accepted, otherwise the hypothesis should be rejected, and (2) the estimate path coefficient for a relationship is significant at the 0.05 cut-off point (Sig.≤ 0.05).

Table 4
.11 indicates the output of bootstrapping statistics for the mediation model (big data analytics → risk management → crisis management).Developing an integrated framework for risk management in policing: crisis management using big data analytics as a case study ___________________________________________________________________________ Rev. Gest.Soc.Ambient.|Miami | v.18.n.2 | p.1-36 | e06035 | 2024.28

Table 7
The summary of bootstrapping of total and indirect effects