Vol. 4 No. 2 (2025): November

Government websites serve as essential platforms for delivering information and public services online, yet many still face usability challenges. The present study evaluates the user interface of the Tulungagung Regency Government website and provides interface improvement recommendations. The research utilized a heuristic evaluation method involving three evaluators and applied ten heuristic principles to identify usability issues. Data were collected through direct observation, documentation, and user questionnaires. Interface improvement recommendations were designed as a high-fidelity prototype using Figma. To test the effectiveness of the proposed design, usability testing was conducted using the System Usability Scale (SUS) method involving 25 respondents. The study found key issues including inconsistent visual design, disorganized menu structure, and lack of system feedback. The usability testing results indicated a significant increase in usability scores with the redesigned prototype compared to the original version. The research contributes to the development of more user-friendly government websites and can serve as a reference for similar interface evaluations in the public sector.

This study addresses logistical inefficiencies at Bintang Technic, a technical equipment trading startup in Malang Regency, East Java, where manual recording led to delays, inaccurate stock data, and limited transparency, with 10–15 monthly distribution delays and over 15% error rates in transactions. Utilizing a Research and Development (R&D) approach, a web-based information system was developed using Laravel and MySQL to streamline incoming goods management. Through qualitative data collection via observations and interviews, the system was designed with integrated database structures, role-specific dashboards for operations (item management, requests), finance (request reviews, vendor management, POs), and managers (approvals, reports), supported by use case diagrams, DFD, and ERD. Stakeholder feedback refined the design, and Black Box Testing across 21 scenarios confirmed 100% functionality, reducing processing times from 2–3 days to under 1 day and errors from 15% to below 3%. The system enhanced transparency, decision-making, and efficiency through real-time data and automation, overcoming challenges like user resistance and infrastructure limitations via training and technical support. This solution not only meets Bintang Technic’s needs but also offers a scalable model for SMEs, with potential for integration with other business modules.

The rapid digitalization of commerce in Indonesia has positioned Tokopedia as a central marketplace that facilitates large-scale transactions while managing vast amounts of sensitive user data. This reliance on digital infrastructures, however, exposes the platform to escalating cyber threats that jeopardize both operational continuity and consumer trust. This study evaluates Tokopedia’s cybersecurity strategies by applying the Risk Assessment Framework derived from ISO 27001 and ISO 31000. Using a qualitative descriptive design, the research draws exclusively on secondary sources such as peer-reviewed journals, industry reports, and case studies published between 2015 and 2025. The analysis identifies five dominant risks: large-scale data breaches, phishing and identity theft, ransomware attacks, insider threats, and system misconfigurations. Risk assessment results indicate that data breaches pose the most critical threat, with phishing and ransomware classified as medium but persistent risks. Tokopedia has implemented several protective measures, including encryption, multi-factor authentication, e-KYC verification, and privacy policies. Nevertheless, gaps remain in governance, routine audits, and employee awareness, leaving the platform vulnerable to recurring incidents. A comparative analysis with global platforms highlights the importance of proactive governance, systematic risk documentation, and continuous training, areas where Tokopedia is still underdeveloped. The findings underscore that cybersecurity should be recognized not merely as a technical safeguard or financial burden but as a strategic investment essential for resilience, consumer confidence, and sustainable growth in Indonesia’s competitive digital economy.

Heart disease remains a leading cause of mortality worldwide, including in Indonesia, and is often difficult to detect at an early stage. One of the main challenges in the Indonesian healthcare system is the lack of fully digitalized data management and the issue of imbalanced patient datasets, which reduce classification accuracy. This study developed a web-based information system designed to manage patient records and automatically classify heart disease risk. The system was implemented using the CodeIgniter framework with a MySQL database, and applied an Artificial Neural Network (ANN) in combination with the Synthetic Minority Over-sampling Technique (SMOTE) to address data imbalance. A total of 60 secondary patient records were processed through preprocessing, data balancing, model training, and cross-validation. Experimental results demonstrated that the application of SMOTE improved model sensitivity, with performance metrics of 87.4% accuracy, 85.2% precision, 88.6% recall, and an AUC-ROC of 0.94. These findings confirm that integrating ANN and SMOTE into a web-based system enhances classification reliability and supports faster medical decision-making. However, the study also acknowledges certain limitations, including the restricted dataset size and the absence of validation in real clinical environments. Future work should expand the dataset, test the system in healthcare facilities, and compare performance with other algorithms such as Random Forest or SVM to identify the most optimal predictive model.

This study examines the development and implementation of Your Favorite Shirt, a web-based platform designed to modernize thrift shop operations through digital transformation. Many thrift businesses in Indonesia still rely on manual processes or social media, leading to inefficiencies in inventory management, transaction tracking, and reporting. To address these challenges, a prototype-based development method was applied, involving requirement analysis, initial design, prototype construction, user testing, and iterative refinement. The system integrates key features such as multi-role login, product catalogues, transaction management, payment options, and order tracking. Evaluation through functional testing and the PIECES framework demonstrated positive results across performance, efficiency, control, and service dimensions, with an average system response time of under one second. User Acceptance Testing (UAT) with a Likert-scale questionnaire produced a mean score of 1.85, indicating strong user satisfaction. Beyond improving operational efficiency, the platform also supports sustainable fashion practices by curating quality secondhand products and reducing textile waste. However, risk assessment identified several areas requiring improvement, including the absence of a Business Continuity Plan, limited access control mechanisms, and lack of systematic documentation. Future development should focus on integrating social media APIs, implementing product recommendation systems, enhancing layered security, and deploying analytical dashboards to support data-driven decision-making.

This study examines the development and implementation of AcaraKita, a web- and mobile-based digital platform designed to modernize event organizer (EO) services in Indonesia. The system integrates three primary actors—Admin, Admin EO, and Customer—each with distinct yet complementary roles involving vendor management, booking, verification, status tracking, and service reviews. The development process applied the Waterfall model, consisting of requirement analysis, design, implementation, testing, deployment, and maintenance, combined with IT governance evaluation using the COBIT 5 framework to ensure alignment with business objectives. Testing results indicated that all core features operated effectively, with an average response time of less than one second and a user satisfaction score of 4.139 on the Likert scale. The IT governance risk analysis highlighted the need for improvements in documentation, security, and business continuity planning. While the system demonstrates a solid foundation, further enhancements are necessary, including social media API integration, vendor recommendation systems, and analytics dashboards to support decision-making. Overall, AcaraKita strengthens EO digitalization, improves operational efficiency, and fosters service transparency in a sustainable manner.

The development of digital transaction services has led to the widespread use of APIs in payment systems, including payment gateway services such as Midtrans. However, the open access to APIs also increases the risk of cyber attacks, one of which is Distributed Denial of Service (DDoS) which can destabilize the system and reduce user confidence. This research aims to analyze the potential DDoS threats to the Midtrans API and explore the application of rate limiting techniques using Node.js as one of the mitigation measures. The methodology used is a waterfall approach, which includes requirements analysis, system design, implementation, testing, and evaluation. The test design is done through simulating DDoS attacks on API endpoints, both before and after the application of rate limiting, by measuring parameters such as the number of requests, response time, and request success rate. It is hoped that this research can provide a clear picture of the importance of API protection in digital payment systems, and produce a technical approach that can be used as a reference in developing a secure and reliable system. This research is also expected to make practical and theoretical contributions in the field of API security and digital service traffic management.

The advancement of information technology has shifted libraries from traditional physical services to digital platforms. One example is JAKLITERA, a web-based information system developed by the Provincial Library and Archives Office of DKI Jakarta. This study aims to evaluate the performance and service quality of JAKLITERA using the PIECES framework (Performance, Information, Economics, Control, Efficiency, Service). A mixed approach was applied, combining quantitative data from 100 user questionnaires with qualitative insights from observation and administrator interviews. Validity and reliability testing confirmed that all research instruments were appropriate for analysis. Results showed that the Economics dimension achieved the highest mean score (3.95), while Control received the lowest (3.39). Overall, the average score of 3.62 categorized JAKLITERA as “satisfactory,” yet highlighted the need for improvements in security mechanisms and clarity of information. These findings suggest that JAKLITERA functions effectively as a digital literacy platform but requires continuous evaluation, technical optimization, and user-centered enhancements to ensure sustainable service delivery.

The rapid growth of information technology has amplified the demand for fast and reliable internet services, particularly in urban centers such as West Jakarta. This study aims to design a predictive model of customer satisfaction for RumahNet’s Fiber to the Home (FTTH) services by applying the Decision Tree (C4.5) algorithm. A survey of 250 active subscribers was conducted using a Likert-scale questionnaire distributed through Google Forms, capturing perceptions of internet speed, connection stability, pricing, and technical support. The dataset was processed and analyzed using RapidMiner Studio within the Knowledge Discovery in Databases (KDD) framework. Results show that the model achieved an accuracy of 85.33%, precision of 91.93%, recall of 90.47%, and an F1-score of 91.18%. The decision tree revealed that internet speed and connection stability were the most critical determinants of satisfaction, followed by pricing and responsiveness of customer service. These findings suggest that prioritizing technical reliability while maintaining affordability and responsive support is essential for strengthening loyalty and reducing churn. The research demonstrates that Decision Tree modeling not only provides high predictive accuracy but also offers clear interpretability, making it a valuable tool for data-driven decision-making in the ISP sector.

This study examines the role of redundancy protocols in enhancing the reliability of core networks within enterprise infrastructures. The research is motivated by the growing reliance of organizations on network-based services and the increasing risks of financial losses caused by downtime, which are projected to escalate alongside global traffic growth. A simulation-based approach was conducted using Cisco Packet Tracer to compare baseline networks without redundancy against implementations of Hot Standby Router Protocol (HSRP) and Virtual Router Redundancy Protocol (VRRP). The evaluation focused on key performance indicators, including failover time, uptime rate, packet loss, latency, and jitter. The results demonstrate that both protocols successfully provided automatic recovery mechanisms, achieving an average failover time of less than five seconds while maintaining stable connectivity. Further analysis revealed that HSRP excels in failover speed, whereas VRRP offers cross-vendor flexibility with consistent stability. These findings confirm that redundancy protocol selection should align with specific network requirements, yet both HSRP and VRRP have proven effective in minimizing downtime risks and ensuring service continuity in enterprise-scale environments.

The rapid growth of embedded systems demands reliable, efficient, and affordable actuators for smart applications such as home automation and IoT. This study develops a low-cost DC motor based on fundamental electromagnetic principles, finite element analysis (FEA), and Arduino-based PWM control. The motor is constructed from simple materials like copper wire, iron core, and neodymium magnets, with a production cost under five US dollars per unit. Experimental results demonstrate efficiency up to 35% at 165 RPM with a 100-turn coil configuration and a 0.4 Tesla magnetic field. FEA is used to validate the design and identify optimal configurations, while PWM control enables precise speed and torque regulation. The motor is integrated into a smart curtain prototype using MQTT, which automatically adjusts curtain position based on light sensor input, proving the motor’s compatibility with modern automation systems. Experimental findings reveal that increasing coil turns and magnetic field strength significantly improves torque, speed, and mechanical stability. This motor offers a cost-effective solution suitable for educational, research, and commercial applications in resource-constrained environments. The study also opens avenues for future brushless motor development and AI-based adaptive control to further enhance performance. By combining mechanical simplicity with electronic sophistication, this motor presents an optimal alternative for embedded systems prioritizing efficiency and affordability. This approach supports the democratization of automation technology, especially in developing countries and educational institutions with limited budgets.

The rapid expansion of e-commerce in Indonesia has resulted in a surge of unstructured online reviews, especially on platforms such as Shopee. These reviews offer valuable insights into customer satisfaction, product complaints, and purchasing behavior but remain largely underutilized due to their volume and informal language style. This study applies Support Vector Machine (SVM) with Term Frequency–Inverse Document Frequency (TF-IDF) feature extraction to classify reviews of Android smartphones into positive, negative, and neutral categories. Using a dataset of 300 manually annotated reviews from Samsung, Xiaomi, and Oppo official stores, the model achieved an accuracy of 76.67% and demonstrated stable results through 5-fold cross-validation. The negative class showed the highest performance (F1 = 0.86), while the neutral class performed weakest (F1 = 0.62), reflecting challenges posed by mixed opinions and underrepresented samples. Compared with Naïve Bayes and Logistic Regression, the SVM model consistently outperformed both baselines, confirming its suitability for high-dimensional text data and informal Indonesian expressions. The findings highlight SVM’s potential to support automated sentiment monitoring in e-commerce, enabling businesses to identify emerging issues, improve customer service strategies, and leverage positive reviews for marketing. Future research should consider larger and more balanced datasets, techniques for handling imbalanced classes, and integration with deep learning models such as LSTM or BERT to improve performance and generalization.

Conventional household water management often results in inefficiencies, including tank overflow, unexpected shortages, and a lack of awareness about daily consumption. Most existing solutions address these issues only partially—either by monitoring water levels, automating pump control, or recording usage data—without integrating the three into a unified system. To address this gap, this research developed and validated a low-cost Internet of Things (IoT) prototype that combines real-time water-level monitoring, daily consumption measurement, and automatic pump control within a smartphone-connected platform. The system is built on a NodeMCU ESP8266 microcontroller equipped with an HC-SR04 ultrasonic sensor, a YF-S201 flow sensor, and a relay-controlled pump, with data transmitted via Wi-Fi to the Blynk application. Using a Research and Development (R&D) methodology with a prototyping model, the study conducted functional, accuracy, and usability testing. Results show that the prototype achieved reliable performance, with an average error below 2% for both sensors and stable operation during 24-hour trials. Beyond technical validation, the system demonstrated its potential as an eco-feedback tool by providing clear consumption data that can encourage more sustainable water use at the household level.

This study presents the development and evaluation of a secure and efficient real-time face recognition system for school attendance, integrating cancelable biometrics with cryptographic hashing. A total of 115 face samples were collected from students and teachers under diverse lighting, pose, and expression conditions. Images were pre-processed using Contrast Limited Adaptive Histogram Equalization (CLAHE) and Gamma Correction, followed by feature extraction with ResNet-128D, key-based random projection, binarization into 128-bit templates, and SHA-256 hashing. Evaluation results demonstrated an accuracy of 86.09%, precision of 100%, recall of 86.09%, and F1-score of 92.52%, with an average latency of 281.71 ms, remaining well below the operational threshold of 500 ms. Offline pre-processing improved the F1-Score by 7.50% on large datasets and 7.28% on smaller datasets without sacrificing processing speed. From a security perspective, the system achieved zero false acceptances (FAR = 0%) and allowed template regeneration when compromised, reinforcing privacy preservation. These findings validate the feasibility of combining cancelable biometrics with cryptographic hashing to balance accuracy, speed, and security in practical attendance systems. The research underscores its broader applicability to access control and public security, while future work should emphasize adaptive pre-processing, diverse hardware validation, and hardware acceleration for robust real-time deployment.

This study examines the effectiveness of classical computer vision methods for modern waste classification by combining Histogram of Oriented Gradients (HOG) for feature extraction with Support Vector Machine (SVM) for classification. The TrashNet dataset, consisting of five categories—cardboard, glass, metal, paper, and plastic—was used as the primary benchmark. To address data limitations and improve generalization, augmentation techniques such as random rotations, horizontal flipping, and brightness adjustments were applied. Hyperparameter optimization was further conducted using GridSearchCV with the RBF kernel to determine the most effective configuration. The optimized model achieved an accuracy of 84.36%, representing a substantial improvement from the 60% baseline. These findings confirm that non-deep learning approaches remain relevant and can serve as computationally efficient alternatives to CNNs, which typically require GPUs and extensive training time. Challenges persist in classifying reflective materials such as glass and metal, where HOG descriptors are less effective. Future work should integrate complementary descriptors, including color and texture-based features, to enhance robustness and scalability. Overall, the study demonstrates that an optimized HOG-SVM pipeline offers a practical, resource-efficient solution for automated waste classification, with strong potential to support sustainable waste management in real-world applications.

This study investigates public sentiment toward the first 100 days of the Prabowo–Gibran administration by analyzing opinions expressed on X (formerly Twitter) using machine learning approaches. A total of 431 valid tweets were collected, preprocessed, and manually labeled into positive and negative categories. The results reveal that 62% of public sentiment was negative, while 38% was positive, indicating widespread public criticism during the administration’s early period. Two algorithms, Naïve Bayes and K-Nearest Neighbor (KNN), were applied to classify sentiment. The Naïve Bayes model achieved superior performance, with an accuracy of 97.22%, compared to KNN’s 62.65%. The probabilistic nature of Naïve Bayes allowed it to manage high-dimensional, imbalanced textual data effectively, while KNN suffered from the “curse of dimensionality” and class bias. These findings demonstrate that Naïve Bayes remains a reliable and computationally efficient model for political sentiment analysis in the Indonesian digital context. Despite its strengths, this study acknowledges limitations in manual labeling and linguistic nuances such as sarcasm and irony. Future research is encouraged to integrate deep learning architectures like LSTM or BERT and adopt aspect-based sentiment analysis to capture more contextual insights from political discourse.

This study investigates the impact of Game-Based Learning (GBL) on students’ learning interest and examines the effectiveness of the Naive Bayes algorithm in predicting engagement levels among junior high school students. Using a quasi-experimental quantitative design, data were collected from fifty seventh-grade students at SMP BQ Boarding School through pre-test and post-test questionnaires administered before and after a four-week GBL intervention. Statistical analysis revealed a significant increase in learning interest, with mean scores rising from 2.85 to 4.10 (t(49) = –10.24, p < 0.001), confirming the positive influence of GBL in promoting motivation and participation. The Naive Bayes classification model achieved an accuracy rate of 90%, with precision and recall values of 0.92 and 0.95 for the high-interest category, respectively. These results demonstrate that GBL effectively transforms classroom dynamics into interactive learning experiences while the Naive Bayes model reliably identifies students’ motivational levels. The combination of pedagogical innovation and predictive analytics presents a practical framework for educators to design adaptive interventions and data-informed teaching strategies. This study underscores the importance of integrating artificial intelligence and game-based methods in education to enhance engagement, motivation, and learning outcomes in the digital era.

This study presents the design, implementation, and validation of a cloud security architecture on Amazon Web Services (AWS) that integrates Defense in Depth strategies with Identity and Access Management (IAM) Best Practices, tailored for small and medium-sized enterprises (SMEs). Using the AWS Free Tier, an experimental cloud infrastructure was constructed to evaluate the effectiveness of multi-layered protection encompassing network segmentation, least-privilege access control, and real-time monitoring. The architecture employed a segmented Virtual Private Cloud (VPC) with public and private subnets, controlled by Security Groups (SGs) and Network Access Control Lists (NACLs), while IAM policies and Multi-Factor Authentication (MFA) enforced identity-level security. Centralized monitoring through CloudTrail and CloudWatch enabled anomaly detection and event logging with high accuracy. Test results showed that unauthorized access was effectively blocked, suspicious activities were detected promptly, and all administrative actions were recorded reliably. The findings indicate that combining layered network defenses and IAM governance significantly enhances the resilience, visibility, and security posture of SMEs adopting AWS environments. Beyond its technical effectiveness, the model offers scalability, auditability, and cost-efficiency—demonstrating that enterprise-grade protection can be achieved even within the resource constraints of SMEs. Future work is encouraged to integrate automation and advanced AWS tools such as GuardDuty and Config to strengthen real-world adaptability and compliance.

In the evolving retail landscape, data-driven decision-making has become essential for understanding customer behavior and predicting sales trends. This study integrates clustering and classification techniques to analyze retail sales data comprising 1,000 transactions obtained from Kaggle. Using the K-Means algorithm, three optimal customer clusters were identified through the Elbow Method, achieving an average within-centroid distance of 25,272.635 and a Davies–Bouldin Index of 0.443, indicating clear cluster separation. The subsequent classification phase compared the predictive performance of three algorithms—Naïve Bayes, Decision Tree, and Random Forest—on 70:30 training-to-testing data partitions. The Naïve Bayes algorithm attained 94.67% accuracy, while both Decision Tree and Random Forest achieved perfect classification accuracy of 100%. These findings highlight the robustness and adaptability of tree-based models for complex retail datasets, outperforming probabilistic methods in terms of accuracy and generalization. The results suggest that the integration of clustering and classification provides retailers with a powerful analytical framework for identifying high-value customer segments, optimizing marketing strategies, and enhancing inventory management. Despite achieving strong outcomes, the study acknowledges dataset limitations and recommends future research involving larger and more diverse datasets, as well as additional features, to expand model scalability and predictive precision.

This study investigates student satisfaction toward Scout extracurricular activities at SDN Pondok Bambu 10 and 11 by applying the C4.5 algorithm within the CRISP-DM framework. Data were collected from 210 students through questionnaires and interviews to evaluate perceptions of program quality, mentor support, and social interaction. The C4.5 model achieved an accuracy rate of 99.52%, effectively identifying key determinants of student satisfaction. Results indicate that program quality, mentor support, and peer interaction are the most influential factors shaping students’ experiences. The decision tree produced interpretable rules that help educators understand satisfaction patterns and make data-driven improvements to program design. Compared with other methods such as SVM and Random Forest, C4.5 provides clearer interpretability while maintaining high predictive precision. The study further recommends integrating the model into a web-based information system to enable continuous monitoring and evaluation of extracurricular activities. These findings highlight the potential of data mining techniques to strengthen decision-making in education and to create a more adaptive, student-centered approach to extracurricular management.

This study investigates public sentiment toward Joko Anwar’s 2025 film Pengepungan di Bukit Duri using computational text analysis on 583 Instagram comments. The research applies the Naïve Bayes algorithm combined with TF-IDF weighting to classify opinions into positive and negative sentiments. Data were collected through web scraping of public Instagram posts related to the film and processed through several stages including data cleaning, manual labeling, text preprocessing, and probabilistic classification. The results reveal that 72.9% of the comments express positive sentiment, while 27.1% are negative, indicating strong audience appreciation for the film’s narrative quality and social themes. The model achieved an accuracy of 83.67%, with a precision of 87.13%, recall of 91.04%, and F1-score of 89.04% for positive sentiment. These findings confirm that the Naïve Bayes approach is effective for analyzing short, informal Indonesian-language texts on social media. Practically, the results provide valuable insights for filmmakers and cultural analysts in understanding audience perceptions, managing digital reputation, and designing sentiment-based marketing strategies. Future research is recommended to employ hybrid models and multi-platform datasets to enhance sentiment detection, particularly for nuanced or negative expressions.

This study examines public sentiment toward the dominance of QRIS and GPN compared to Mastercard and Visa, using data collected from Instagram comments. Employing the Knowledge Discovery in Databases (KDD) methodology and the Naïve Bayes Classifier, the research analyzed 820 comments retrieved through automated scraping and processed using text mining techniques such as case folding, tokenization, stopword removal, stemming, and TF-IDF transformation. The model achieved an accuracy of 84.27%, a precision of 86.09%, a recall of 94.7%, and an F1-score of 90.21%, indicating strong reliability in identifying sentiment polarity. The analysis revealed that 76.5% of the comments expressed positive sentiment, highlighting users’ appreciation of QRIS and GPN for their convenience, speed, and accessibility across both micro and macro-scale transactions. Negative comments, representing 23.5%, centered on concerns about connectivity, data security, and trust in financial governance. These findings suggest that while QRIS and GPN have been widely embraced as efficient digital payment solutions, there remains a need for improved infrastructure, user education, and data protection. The study demonstrates the effectiveness of the Naïve Bayes algorithm for large-scale sentiment analysis in multilingual online environments and provides empirical insights for policymakers to strengthen Indonesia’s digital payment ecosystem.

This research introduces a mobile-based system for real-time identification of ornamental rose varieties using the YOLOv8 deep learning algorithm. Motivated by the growing interest in ornamental plants during the COVID-19 pandemic and the high penetration of smartphone users in Indonesia, the study aims to create an efficient and accessible flower recognition tool. A dataset of 813 labeled rose images—red, white, yellow, orange, and pink—was collected from the Roboflow platform and processed using data augmentation techniques to improve model generalization. The YOLOv8 model was trained with 100 epochs, a batch size of 16, and the SGD optimizer, then converted to TensorFlow Lite for mobile deployment through the Flutter framework. Experimental results achieved a mean average precision (mAP50–95) of 0.581, with strong detection performance across most classes. The system successfully operated offline, delivering real-time classification accuracy despite dataset imbalance, particularly in the orange rose class. These findings demonstrate that YOLOv8 can be effectively adapted for mobile horticultural applications, offering practical benefits for flower sorting, crop management, and educational use. Future studies are recommended to expand dataset diversity, enhance environmental testing, and explore cloud-based integration for scalable deployment.

This study analyzes customer sentiment toward Alisa Batik Solo’s TikTok e-commerce using the Naïve Bayes algorithm. A total of 626 customer comments were collected through manual data crawling, cleaned, labeled, and processed using text preprocessing techniques including cleaning, case folding, tokenization, stopword removal, and stemming. The processed data were then transformed using TF-IDF feature weighting and classified with Naïve Bayes to determine the polarity of customer opinions. The evaluation results showed an accuracy of 90.85%, precision of 98.29% for positive sentiment, recall of 95.24%, and an F1-score of 96.72%, indicating that the model performs effectively in classifying Indonesian short-text reviews. The findings reveal that 75.6% of the comments expressed positive sentiment, while 24.4% reflected negative opinions, demonstrating a strong level of customer satisfaction and trust in Alisa Batik Solo’s products and online engagement strategy. This research confirms that the integration of Naïve Bayes with TF-IDF preprocessing provides reliable results in social media sentiment analysis and can serve as a strategic tool for e-commerce businesses to enhance marketing decisions and service quality

This study presents the design, development, and evaluation of an automatic temperature control system for broiler chicken coops using NodeMCU and DHT11 sensors integrated within an Internet of Things (IoT) framework. The system was designed to maintain coop temperature stability automatically, minimizing manual intervention and optimizing environmental conditions for broiler productivity. Using a Research and Development (R&D) approach, the system was constructed with hardware components including NodeMCU ESP8266, DHT11 sensor, servo motor, relay, lamp, and cooling fan, while the software utilized Arduino IDE, Python, and Telegram Bot API for real-time monitoring. The seven-day experimental testing, with thirty readings per day, demonstrated that the system maintained an average temperature of 27.6°C (±0.8°C), achieving 98.5% accuracy compared to manual thermometers, with an average error of 0.65%. The actuators exhibited an average response time of 1.8 seconds, ensuring quick adaptation to environmental changes and preventing heat stress in broilers. The automation reduced manual monitoring time by 80% and inspection frequency by 83%, while lowering energy consumption by approximately 40% through temperature-based device activation. These results confirm that low-cost IoT automation enhances environmental stability, animal welfare, and operational efficiency, aligning with the global trend toward precision livestock farming. Future improvements should focus on integrating multi-node systems, adaptive control algorithms, and humidity regulation to expand scalability, reliability, and sustainability in poultry management.

This research developed an Internet of Things (IoT)–based automatic watering system designed to support the maintenance of Aglaonema ornamental plants. The system utilizes a NodeMCU ESP8266 microcontroller connected to a soil-moisture sensor and a DHT11 temperature sensor, all integrated with the Blynk mobile platform for real-time monitoring and remote control. By combining sensor feedback with automated logic, the system activates irrigation when soil moisture falls below 40 percent and stops it when levels exceed 60 percent. Laboratory and field testing confirmed that the prototype operated reliably within a six-meter Wi-Fi range, maintaining stable communication and accurate sensor readings. The device reduced water consumption by about 37 percent compared with manual watering and provided timely notifications through both the LCD display and the Blynk application. Although system performance depends on Wi-Fi connectivity and lacks backup power, its overall operation demonstrates that IoT-based automation can significantly improve water efficiency and convenience for small-scale plant owners. Future enhancement may include solar-powered modules, additional sensors, and cloud data storage to strengthen system reliability and expand smart-agriculture applications.

This study examines the implementation of the Simple Network Management Protocol (SNMP) integrated with the Cacti monitoring platform to evaluate enterprise network performance within a simulated environment using PNETLab. A quantitative approach was applied through continuous data collection and measurement of key performance indicators such as throughput, packet loss, delay, and availability. The experiment utilized virtual Mikrotik routers connected to an Ubuntu-based Cacti server configured for SNMP polling and RRDTool data storage. Real-time visualization enabled efficient tracking of network behavior and early detection of anomalies. The results showed that under normal conditions, the network achieved stable performance with throughput between 70–90% of link capacity, zero packet loss, latency below 150 milliseconds, and availability above 99%, meeting ITU-T/TIPHON Quality of Service (QoS) standards. When faults were simulated, the system accurately detected and displayed traffic interruptions, allowing rapid identification and resolution of network issues. Compared with other monitoring tools such as Zabbix and Nagios, the SNMP–Cacti integration proved simpler to configure while maintaining analytical precision and reliability. These findings confirm that Cacti, supported by SNMP, provides an efficient, scalable, and low-overhead solution for enterprise network monitoring. Future development may incorporate SNMPv3 for enhanced security and automated alert systems or predictive analytics to improve responsiveness and proactive maintenance in larger infrastructures.

The development of the Internet of Things (IoT) has driven the implementation of smart buildings that rely on real-time sensor data collection and analysis. However, cloud computing-based systems often face problems of high latency and large network loads. This research implements an edge computing architecture to optimize sensor data collection in smart buildings. A prototype was built using edge nodes (Raspberry Pi) that process data from temperature, humidity, light, and motion sensors locally before sending it to the cloud. Test results show that edge computing can reduce latency by up to 45% and reduce data traffic to the cloud by 60%, while also improving the energy efficiency of sensor devices. Thus, edge computing has been proven to effectively improve the performance and efficiency of data collection systems in smart buildings

This study aims to describe how the Comsoft Aeronautical Data Access System (CADAS) is used in the NOTAM workflow at Palembang Aeronautical Information Service and how operators behave when the system is not stable. This research was conducted with a qualitative case study approach, which combined interviews, participatory observations, and document analysis to capture daily operational practices and problems that occurred in real conditions. The results of this study indicated that CADAS is still the main system for NOTAM drafting, verification, and transmission. However, its performance is often disturbed by irregular system errors such as sudden shutdowns, delays in data processing, and intermittent failures to send or receive raw NOTAM information. These disturbances create obstacles in fulfilling time-sensitive procedural requirements especially the ten-minute verification standard as stated in SOP.012. In case of system unresponsiveness, operators switch to alternative procedures through email distribution, Web Flight Plan submission, and manual logbook documentation so as not to cause delays; however, these manual steps require continuous follow-up to avoid gaps in recording. In general, this study underlines the importance of increasing CADAS reliability and ensuring that backup procedures are consistently applied so that continuity can be maintained even if there is an interruption in the system.

Low-light conditions continue to challenge real-time face recognition because dim illumination often produces noisy and low-contrast images that weaken the quality of feature extraction. This study investigates how different preprocessing strategies influence the performance of a dlib-ResNet–based recognition system under such conditions. Two reference dataset sizes—33 and 1000 images—were used to observe how reference variation affects embedding stability. Enhancement was applied either offline to the reference dataset or in real time to incoming video frames, and both approaches were also tested in combination. The experiments show that offline preprocessing provides the most reliable improvement. Enhancing reference images raised the F1-Score by 7.28% (small dataset) and 7.50% (large dataset) without reducing processing speed, indicating that clearer embeddings at registration contribute to more stable matching. Real-time preprocessing, however, produced inconsistent results. While slight gains appeared in specific cases, the added computation and occasional distortion of facial structure reduced accuracy in other scenarios. The combined method produced the weakest performance, with the large dataset showing a 33.71% decline, suggesting that excessive modification disrupts structural consistency between reference and test images. Overall, the results highlight the importance of maintaining coherent facial features rather than applying aggressive adjustments to every frame. Offline enhancement is the most practical strategy for low-light deployments, whereas real-time enhancement should be used selectively. Future work may explore adaptive illumination adjustment capable of tuning enhancement parameters automatically to match varying lighting conditions.