Usage Data Research Articles

A review of strategic decision-making in marketing through big data and analytics

This review paper delves into the transformative impact of big data and analytics on strategic marketing decision-making. Examining the integration of vast datasets and analytical tools in marketing strategies highlights how these technological advancements enable a deeper understanding of customer behavior, enhance product development, and provide a competitive edge. The review underscores the importance of data-driven insights in formulating personalized marketing strategies and the critical role of analytics in predictive and prescriptive decision-making. It addresses the challenges and ethical considerations associated with big data usage, emphasizing the need for robust data governance and ethical practices. The paper suggests future research directions, focusing on emerging technologies and methodologies that could further influence strategic marketing decisions.

Big data and AI in employment: The dual challenge of workforce replacement and protecting customer privacy in biometric data usage

The integration of Artificial Intelligence (AI) and Big Data is ushering in profound transformations across various industries, with biometric data usage standing out due to its deep implications for workforce dynamics and customer privacy. This review article critically examines the dual challenges presented by AI-driven automation and the extensive use of biometric data analytics, focusing on the resultant job displacement and escalating privacy concerns. Biometric technologies such as facial recognition, fingerprint identification, and voice analysis are increasingly deployed across sectors including finance, healthcare, and retail. These technologies aim to enhance security measures, improve user experience, and optimize operational efficiencies. However, they also bring to light substantial ethical dilemmas, particularly concerning the privacy of individuals and the security of the data being collected. The pervasive collection and analysis of biometric data can lead to invasive surveillance and profiling, exacerbating risks to personal privacy. Moreover, the use of AI in automating tasks that were traditionally performed by human workers is leading to significant shifts in employment structures. While AI can increase efficiency and reduce costs, it also raises the specter of widespread job displacement. This potential for automation-driven unemployment is especially pronounced in sectors that heavily utilize routine, repetitive tasks, posing critical socio-economic challenges. This article also explores the regulatory and technological frameworks currently in place, and those that are needed to address these challenges. The effectiveness of existing data protection laws, such as the General Data Protection Regulation (GDPR) in the European Union, and the California Consumer Privacy Act (CCPA) in the United States, is assessed in the context of AI and biometric data. We discuss the role of policy in shaping the ethical use of AI and protecting workers, along with the technological safeguards that could be implemented to secure biometric data and ensure privacy. By synthesizing insights from recent research, case studies, and expert analyses, this article provides a comprehensive overview of how AI and Big Data are reshaping the landscape of work and privacy. It critically discusses the need for a balanced approach that harnesses the benefits of technological advancements while safeguarding individual rights and employment security.

The Roles of Big Data for Supporting E-Government Application and Usage in Public Sector, Dubai

The Roles of Big Data for Supporting E-Government Application and Usage in Public Sector, Dubai

Rollout of a statewide Australian telestroke network including virtual reality training is associated with improved hyperacute stroke workflow metrics and thrombolysis rate

BackgroundTelestroke networks aim to address variability in both quality and access to stroke care in rural areas, by providing remote access to expert stroke neurologists. Implementation of telestroke requires adaptation of workflow processes and education. We previously developed virtual reality (VR) workflow training and documented acceptability, utility and feasibility. The effects on acute stroke treatment metrics have not been previously described.AimsThe overall aim was to improve hyperacute stroke metrics and shorten the time-to-reperfusion therapy administration in rural settings.MethodsThis study applies a natural experiment approach, collecting stroke metric data during transition from a pre-existing pilot to a statewide telestroke service at five rural hospitals. Pre- and post-intervention data included baseline patient demographics and assessment, diagnosis, and treatment delivery metrics. The primary study outcome was door-to-decision time (thrombolysis and endovascular thrombectomy). Secondary outcomes included door-to-computerized tomography time, door-to-thrombolysis time and proportion of patients receiving thrombolysis or thrombectomy treatment. Usage data relating to the VR stroke workflow training of interprofessional healthcare professionals was automatically captured via Wi-Fi. Statistical comparisons of clinical metrics between the pre- and post-intervention time periods, defined as the timeframes before and after VR deployment, were performed.ResultsA total of 2,683 patients were included (April 2013–December 2022); 1910 pre- and 773 post-intervention. All acute stroke time metrics significantly improved post-intervention. The primary outcome, door-to-decision time, decreased from 80 min [56–118] to 54 min [40–76; P < 0.001]. Secondary outcomes also improved, including door-to-thrombolysis time (90 min [68–114] vs. 68.5 min [54–90]; P < 0.001) and proportion of patients thrombolysed (11 vs. 16%; P < 0.001). The proportion of patients transferred for thrombectomy was unchanged (6 vs. 7%; P = 0.69). Seventy VR sessions totaling 15 h 39 min of training time were logged. VR training usage varied across sites (3–31 sessions per site).ConclusionsDelivery of a multi-factorial intervention including infrastructure, funding, education and training (with VR workflow training) as part of a state-wide telestroke rollout was associated with improved acute stroke treatment metrics. Additional work is required to identify the contribution of each intervention component on clinical outcomes and to increase training uptake and sustainment.

Improving Electricity Theft Detection Using Electricity Information Collection System and Customers’ Consumption Patterns

Electricity theft detection (ETD) techniques employed to identify fraudulent consumers often fail to accurately pinpoint electricity thieves in real time. The patterns associated with electricity use are leveraged to identify anomalies indicative of electricity theft. However, challenges in the benchmark ETD include overfitting and a high incidence of false positives (FPs) resulting from incorrect usage patterns formed by considering only electricity consumption patterns without accounting for external factors that contribute to variations in normal consumption patterns. Further investigation is required to precisely detect instances of electricity theft, thereby minimizing nontechnical losses and forecasting future electricity demand to maintain a stable supply. This study employs a master energy meter located on the transformer side to monitor the amount of energy distributed to the region. Zones with a high likelihood of energy theft are detected by comparing the sum of readings from all the smart meters with the readings from the master energy meter installed on the HV of the substation transformer within the same electric feeder. Ensemble XGBoost machine-learning algorithm and K-Means algorithm are used for the classification of malicious and nonmalicious samples and grouping similar types of consumers together, respectively. This makes the proposed model resistant to false-positive rates caused by changes in usage patterns that aren’t done on purpose. Furthermore, energy thieves are identified by detecting anomalies in consumption behavior while maintaining constant internal and external environmental variables. This novel model proposed here mitigates the FP rate found in the present research of electricity usage data. Our approach outperforms support vector machines, convolution neural network, and logistic regression in simulations. Precision, F1-score, recall, Matthews Correlation Coefficient, Receiver Operating Characteristics (ROC)-Area Under The Curve (AUC), and Precision Recal (PR)-Area Under The Curve (AUC) validate our model. The simulation results show that the proposed K-Means-LSTM-XGBoost model improved the classifier’s F1-score, precision, and recall to 93.75%, 95.16%, and 92.38%, respectively. Our model classifies huge time series data with high precision and can be utilized by the utility for real time electricity theft detection.

Reporting cybersecurity to stakeholders: A review of CSRD and the EU cyber legal framework

The purpose of this article is to explain, through a doctrinal review of the EU sustainability and cybersecurity legal framework, how the cybersecurity obligations contribute to the cybersecurity content and quality of the sustainability reporting. Previous studies are limited to voluntary cybersecurity disclosure in annual reports because they date back to before the adoption of CSRD. The CSRD harmonized sustainability reporting in the EU and introduced the ESRS, the standards for sustainability disclosure. As stated in the ESRS S4, the sustainability reporting should now address how the company manages the risks linked to data usage and data collection. Therefore, cybersecurity measures must be included in the sustainability report. This information can be used by stakeholders to assess the risk appetite and the potential long-term profitability of the company. The cybersecurity measures adopted by companies must comply with the cybersecurity obligations of the EU legal framework. While it is out of the scope of these cybersecurity obligations to inform ex ante the stakeholders of a company of how the company is managing cyber risks, these same obligations can improve the quality and content of the sustainability discourse.

The pivotal role of data harmonization in revolutionizing global healthcare: a framework and a case study

Aim: Data harmonization standardizes healthcare information, enhancing accessibility and interoperability, which is crucial for improving patient outcomes and driving medical research and innovation. It enables precise diagnoses and personalized treatments, and boosts AI model efficiency. However, significant challenges such as ethical concerns, technical barriers in the data lifecycle, AI biases, and varied regional regulations impede progress, underscoring the need for solutions like adopting universal standards such as HL7 FHIR, where the lack of generalized harmonization efforts is significant. Methods: We propose an advanced, holistic framework that utilizes FAIR-compliant reference ontologies (based on the FAIRplus and FAIR CookBook criteria) to make data findable, accessible, interoperable, and reusable enriched with terminologies from OHDSI (Observational Health Data Sciences and Informatics) vocabularies and word embeddings to identify lexical and conceptual overlaps across heterogeneous data models. Results: The proposed approach was applied to autoimmune diseases, cardiovascular diseases, and mental disorders using unstructured data from EU cohorts involving 7,551 patients with primary Sjogren’s Syndrome, 25,000 patients with cardiovascular diseases, and 3,500 patients with depression and anxiety. Metadata from these datasets were structured into dictionaries and linked with three newly developed reference ontologies (ROPSS, ROCVD, and ROMD), which are accessible on GitHub. These ontologies facilitated data interoperability across different systems and helped identify common terminologies with high precision within each domain. Conclusion: Through the proposed framework, we aim to urge the adoption of data harmonization as a priority, emphasizing the need for global cooperation, investment in technology and infrastructure, and adherence to ethical data usage practices toward a more efficient and patient-centered global healthcare system.

The Development of a European Multidisciplinary Cleft Lip and Palate Registry by the European Reference Network CRANIO: Experiences, Barriers, And Facilitators.

The European Reference Network for Rare Craniofacial Aanomalies and Ear-Nose-Throat disorders aims to improve care for patients with such afflictions, including cleft lip and palate (CL/P) across Europe. Cleft treatment remains varied throughout European centers, inhibiting meaningful comparison of treatment outcomes. To overcome these issues, a European-wide common CL/P dataset and registry was developed, facilitating standardized treatment endpoints and outcome measures for international comparison and benchmarking of CL/P centers. Questionnaires and semi-structured interviews were used to determine the set-up of the registry. Previous CL/P initiatives were analyzed to create an initial dataset, refined through consensus meetings. In total, 87 cleft specialists working in specialized CL/P centers from 16 European nations participated. Consensus on a common dataset was reached. A "Level 1" dataset, with mandatory clinical and patient-reported outcome measures, and "Level 2" dataset with additional outcome measures. Finally, 2 dashboards were developed for data dissemination. The development of the European CL/P common dataset and registry tackled challenges with resource disparities, variations in specialists within CL/P teams, regulatory differences in patient data usage, patient-reported outcome measures availability in European languages, and use of assessment tools. This study described the successful development of the European Reference Network for Rare Craniofacial Aanomalies and Ear-Nose-Throat disorders CL/P common dataset and registry. This achievement will help improve patient care and outcomes for patients with CL/P in Europe. Furthermore, this study provides useful information for initiatives with similar aims.

DNA Cryptography for Enhanced Data Storage Security in Cloud Environments

Despite the persistent security challenges inherent in cloud systems, a distributed cloud environment necessitates an access control model that is contextually aware to effectively manage these challenges. This model should incorporate a role activation process based on the user's contextual information. Within this role activation process, the rationale behind data collection and usage is disclosed, enabling administrators to establish context-based policies. Consequently, role permissions are dynamically activated based on the association of roles with context. To mitigate complications in the role-based access control model, users are categorized into classes or groups, each with its own access control standards. Access to specific resources is determined by the user's identity upon request. Traditional access control models often fall short in cloud environments due to their inability to address all aspects of the diverse entities, resources, and users present. In the proposed access control system with perception reasoning, entities are expanded using Extensible Access Control Markup Language (XACML), while a trust module monitors user behavior dynamically, detecting and restricting malicious users attempting illegal data access. This includes assigning an identity tag to malicious users, which involves task and data classification along with database tagging.

Evaluation of different landing pages on behavioural engagement with the CARA dashboard: A user research protocol

BackgroundCARA set out to develop a data-visualisation platform to facilitate general practitioners to develop a deeper understanding of their patient population, disease management and prescribing through dashboards. To support the continued use and sustainability of the CARA dashboards, dashboard performance and user engagement have to be optimised. User research places people at the centre of the design process and aims to evaluate the needs, behaviours and attitudes of users to inform the design, development and impact of a product.ObjectiveTo explore how different initial key messages impact the level of behavioural engagement with a CARA dashboard.MethodsParticipating general practices can upload their practice data for analysis and visualisation in CARA dashboards. Practices will be randomised to one of three different initial landing pages: the full dashboard or one of two key messages: a between comparison (their practice prescribing with the average of all other practices) or within comparison (with practice data of the same month the previous year) with subsequent continuation to the full dashboard. Analysis will determine which of the three landing pages encourages user interaction, as measured by the number of ‘clicks’, ‘viewings’ and ‘sessions’. Dashboard usage data will be collected through Google analytics.DiscussionThis study will provide evidence of behavioural engagement and its metrics during the implementation of the CARA dashboards to optimise and sustain interaction.Trial registrationISRCTN32783644 (Registration date: 02/01/2024).

Deep Anomaly Detection Framework Utilizing Federated Learning for Electricity Theft Zero-Day Cyberattacks.

Smart power grids suffer from electricity theft cyber-attacks, where malicious consumers compromise their smart meters (SMs) to downscale the reported electricity consumption readings. This problem costs electric utility companies worldwide considerable financial burdens and threatens power grid stability. Therefore, several machine learning (ML)-based solutions have been proposed to detect electricity theft; however, they have limitations. First, most existing works employ supervised learning that requires the availability of labeled datasets of benign and malicious electricity usage samples. Unfortunately, this approach is not practical due to the scarcity of real malicious electricity usage samples. Moreover, training a supervised detector on specific cyberattack scenarios results in a robust detector against those attacks, but it might fail to detect new attack scenarios. Second, although a few works investigated anomaly detectors for electricity theft, none of the existing works addressed consumers' privacy. To address these limitations, in this paper, we propose a comprehensive federated learning (FL)-based deep anomaly detection framework tailored for practical, reliable, and privacy-preserving energy theft detection. In our proposed framework, consumers train local deep autoencoder-based detectors on their private electricity usage data and only share their trained detectors' parameters with an EUC aggregation server to iteratively build a global anomaly detector. Our extensive experimental results not only demonstrate the superior performance of our anomaly detector compared to the supervised detectors but also the capability of our proposed FL-based anomaly detector to accurately detect zero-day attacks of electricity theft while preserving consumers' privacy.

Process evaluation of a Structured E-parenting Support (STEPS) in the OPTIMA randomised controlled trial: a protocol.

Structured E-parenting Support (STEPS) is a digital application (app) designed to help parents manage behaviour of their children who are referred to mental health services and are waiting for an assessment or treatment. STEPS is currently being evaluated in the Online Parent Training for the Initial Management of Attention-Deficit/Hyperactivity Disorder randomised controlled trial. Alongside the examination of STEPS' clinical and cost-effectiveness, we are conducting a process evaluation to better understand the contextual factors that may influence study outcomes. The purpose of this protocol is to describe the aims, objectives and methodology of the process evaluation prior to it taking place to add to the fidelity and rigour of the trial process and outcomes. Our goal is to adapt STEPS to optimise its benefits in future applications. In line with the Medical Research Council guidelines for evaluating complex interventions, the process evaluation will adopt a mixed method design using qualitative data collected from clinicians and parent interviews and app usage data from participants assigned to the intervention arm. Qualitative data from semistructured interviews and free text box responses included in trial questionnaires will be analysed thematically using framework analysis to better understand how parents use STEPS, how it works and key factors that could aid or hinder its effective implementation in routine clinical practice. The application for ethical approval for the study was submitted to the North West-Liverpool Central Research Ethics Committee and received a favourable opinion on further information on 26 November 2021, reference number 21/NW/0319. The process evaluation aims to explore how a digital app might support parents in managing their child's behaviour. Implications for policy and research will be explored and the clinical implications of offering the app to a wider audience to address the lack of support to parents as highlighted in this paper. We plan to publish findings in international, peer-reviewed journals as well as present at conferences. The trial has been prospectively registered on 18 November 2021; ISRCTN816523503. https://www.isrctn.com/ISRCTN16523503.

Nonlinear Analysis and Topological Approaches towards a Deep Intelligent Framework for Privacy Assurance of Autonomous IoT Systems

The proliferation of autonomous Internet of Things (IoT) systems powered by deep learning and artificial intelligence has ushered in a new era of data-driven convenience and automation. However, this innovation comes hand in hand with heightened concerns regarding data privacy. This paper presents a comprehensive framework for Privacy Assurance in Autonomous IoT Systems (PAIS), which amalgamates cutting-edge technologies and best practices to safeguard individual privacy in the era of pervasive connectivity and autonomous decision-making. The PAIS framework comprises multifaceted strategies to address privacy challenges in autonomous IoT ecosystems. It leverages advanced encryption techniques, robust access control mechanisms, and anonymization protocols to ensure data confidentiality. Moreover, differential privacy mechanisms are deployed to protect the identities of individuals within data streams. An innovative aspect of PAIS is the integration of AI-driven privacy monitoring, which constantly evaluates data for potential breaches and triggers immediate responses when anomalies are detected. Ensuring regulatory compliance is a paramount facet of the PAIS framework, as it aligns with evolving data protection regulations globally. Users are afforded control and transparency through intuitive interfaces, enabling them to manage their data usage preferences effectively. The ethical implications of AI in privacy preservation are also examined within the framework, emphasizing the importance of fairness and bias mitigation. PAIS promotes a privacy-by-design approach, where privacy considerations are integral to the inception and development of IoT systems. Regular risk assessments are performed to identify potential privacy vulnerabilities, ensuring that the framework adapts to emerging threats. Education and training programs are provided to stakeholders to foster awareness and adherence to privacy best practices.

The design of a scientific data management system based on DOMAS at CSNS-II (preliminary stage)

At the second stage of China Spallation Neutron Source (CSNS-II), it is predicted that 2 PB raw experimental data will be produced annually from twenty instruments. Scientific computing puts forward higher requirements for data sharing, utilization, retrieval, analysis efficiency, and security. However, the existing data management system (DMS) based on ICAT has several limitations including poor scalability of metadata database, imperfect data-management lifecycle and inflexible API. To ensure the accuracy, usability, scalability and efficiency of CSNS-II experimental data, a new scientific data management system is therefore designed based on the DOMAS framework developed by the Computing Center of IHEP. The data acquisition, transmission, storage and service systems are re-designed and tailored specifically for CSNS-II. Upon its completion, the new DMS will overcome the existing challenges and offer functions such as online display, search functionality and rapid download capabilities for metadata, raw data and analyzed data; flexible and user-friendly authorization; and data lifecycle management. Ultimately, the implementation of the new Data Management System (DMS) is expected to enhance the efficiency of experimental data analysis, propelling CSNS-II to achieve international advanced standards. Furthermore, it aims to reinforce self-reliance and technological strength in the field of science and technology at a high level in China. The development and deployment of the new DMS begin at the end of 2023.

Does derivative usage boost firm value in an economy with controls? Evidence from India

AbstractThe role of derivatives on firm valuation has recently become a subject of intense debate for emerging markets where there is greater asset price volatility, illiquidity, and the absence of common regulations and accounting standards. Given the lack of data availability, disclosure inconsistency and distinctive regulatory requirements for Indian firms, we assess the impact of corporate hedging via derivative use on the performance of BSE-100 non-financial firms over the period 2010–2017. Our results from keywords-based hand-collected derivative usage data show that all types of derivatives, except rate derivatives, have a significant and positive impact on firm value through channels of reduced earnings volatility and increased leverage with tax benefits. Firms with greater leverage tend to benefit from the use of rate derivatives. Still, among those highly leveraged firms, the ones who experience high earnings volatility are unable to benefit from rate derivatives. Besides, hedge accounting benefits firms in dealing with earnings volatility when they have higher derivatives exposure. As capital controls on international flows can limit the use of derivatives, we show that firms with currency risk exposure in terms of long-term foreign currency debt do benefit from rate derivatives. Our results are also robust with a matched sample for derivative users and non-users. Overall, we conclude that easing regulatory controls on derivatives dealings can be beneficial for firms in growing emerging economies.

Semantic-enabled architecture for auditable privacy-preserving data analysis

Small and medium-sized organisations face challenges in acquiring, storing and analysing personal data, particularly sensitive data (e.g., data of medical nature), due to data protection regulations, such as the GDPR in the EU, which stipulates high standards in data protection. Consequently, these organisations often refrain from collecting data centrally, which means losing the potential of data analytics and learning from aggregated user data. To enable organisations to leverage the full-potential of the collected personal data, two main technical challenges need to be addressed: (i) organisations must preserve the privacy of individual users and honour their consent, while (ii) being able to provide data and algorithmic governance, e.g., in the form of audit trails, to increase trust in the result and support reproducibility of the data analysis tasks performed on the collected data. Such an auditable, privacy-preserving data analysis is currently challenging to achieve, as existing methods and tools only offer partial solutions to this problem, e.g., data representation of audit trails and user consent, automatic checking of usage policies or data anonymisation. To the best of our knowledge, there exists no approach providing an integrated architecture for auditable, privacy-preserving data analysis. To address these gaps, as the main contribution of this paper, we propose the WellFort approach, a semantic-enabled architecture for auditable, privacy-preserving data analysis which provides secure storage for users’ sensitive data with explicit consent, and delivers a trusted, auditable analysis environment for executing data analytic processes in a privacy-preserving manner. Additional contributions include the adaptation of Semantic Web technologies as an integral part of the WellFort architecture, and the demonstration of the approach through a feasibility study with a prototype supporting use cases from the medical domain. Our evaluation shows that WellFort enables privacy preserving analysis of data, and collects sufficient information in an automated way to support its auditability at the same time.

Estimated prophylactic dose required to achieve 3% trough as a function of age and concentrate class in multi-country severe WAPPS-Hemo haemophilia patients.

Web-Accessible Population-Pharmacokinetic Service-Haemophilia (WAPPS-Hemo) data are available to study factor-concentrate usage, defined as the required weekly dose to achieve a 3% trough (WD3T), across standard and extended half-life (SHL/EHL) products. To provide baseline usage data including (i) differences across plasma-derived (pdSHL) versus recombinant (rSHL) products, (ii) SHL versus EHL, and (iii) effect of age and positive inhibitor history. PK profiles (n=14,416 patients, 0.3-85.2years) and linear mixed effects models were used to estimate usage versus age, controlling for significant factors, using 95% confidence intervals to perform comparisons across all ages and posthoc tests to assess the differences. Average usage was significantly higher for pdSHL versus rSHL in patients with a positive inhibitor history (PIH; 1.9-2.5 times higher), for SHL versus EHL (4-10 times), and was significantly associated with age. Baseline usage patterns from 2017 to early 2023 provide a benchmark for assessing the impact of emerging technologies in haemophilia.

Analyzing the Energy Usage of a Community and the Benefits of Energy Storage

Understanding the energy usage of a community is crucial for policymaking, energy planning, and achieving sustainable development. The advent of the smart grid has made it feasible to gather fine-grain energy usage data at large-scales, providing us with new opportunities to understand demand patterns at different spatial and temporal scales. In this paper, we conduct a large-scale empirical study of energy usage of 14,849 residential and commercial energy consumers from a small city in the United States. We conduct a wide ranging analysis of energy usage at multiple granularities—citywide, transformer-level, and individual home levels. In doing so, we demonstrate how city-wide smart meter datasets can answer a variety of questions on energy consumption, such as the impact of weather on energy usage. For example, we show that extreme weather events significantly increase energy usage, e.g., by 36% and 11.5% on hot summer and cold winter days, respectively. As another example, we show 19.2% of transformers in the grid get overloaded during peak load periods. Finally, we evaluate the impact of incorporating varying amounts of energy storage within the distribution grid and the impact such deployments will have on the peak demand patterns seen by the grid as well as the ability to reduce overloads seen by distribution transformers during peak periods.

A review of strategic decision-making in marketing through big data and analytics

This review paper delves into the transformative impact of big data and analytics on strategic marketing decision-making. Examining the integration of vast datasets and analytical tools in marketing strategies highlights how these technological advancements enable a deeper understanding of customer behavior, enhance product development, and provide a competitive edge. The review underscores the importance of data-driven insights in formulating personalized marketing strategies and the critical role of analytics in predictive and prescriptive decision-making. It addresses the challenges and ethical considerations associated with big data usage, emphasizing the need for robust data governance and ethical practices. The paper suggests future research directions, focusing on emerging technologies and methodologies that could further influence strategic marketing decisions. Keywords: Big Data, Analytics, Strategic Marketing, Data-Driven Decision-Making, Ethical Considerations, Emerging Technologies.

Efficient Task-driven Video Data Privacy Protection for Smart Camera Surveillance System

As one of the most commonly used AIoT sensors, smart cameras and their supporting services, namely cloud video surveillance (CVS) systems, have brought great convenience to people’s lives. Recent CVS providers use different machine learning techniques to improve their services (regarded as tasks) based on the uploaded video. However, uploading data to the CVS providers may cause severe privacy issues. Existing works that remove privacy information could not achieve a high tradeoff between data usability and privacy, because the importance of information varies with the task. In addition, it is challenging to design a real-time privacy protection mechanism, especially in resource-constrained smart cameras. In this work, we design a task-driven and efficient video privacy protection mechanism for a better tradeoff between privacy and data usability. We use Class Activation Mapping to protect privacy while preserving data usability. To improve the efficiency, we utilize the motion vector and residual matrix produced during video codec. Our work outperforms the region of interest–based methods in data protection while preserving data usability. The attack accuracy drops 70%, while the task accuracy is comparable to those without protection (within ± 4%). The average protection frame rate of the High Definition video can exceed 16 fps+ even on a CPU.