Computer and Science Industrial Engineering (COMASIE) https://forum.upbatam.ac.id/index.php/comasiejournal <p style="text-align: justify;">Comasie Journal is a scientific journal of the students of the Faculty of Engineering and Computer, Putera Batam University, which consists of three groups of Sciences namely information systems, informatics engineering and industrial engineering. The journal will be published six months in one year. This journal contains the scientific work of student research with the specific subjects: Networking, Artificial Intelligence, Image Processing, Expert Systems, Computer Systems, Soft Computing, Web Programming, Mobile Programming, Games Programming, Data Mining, Decision Support Systems, and Ergonomics , Product Design, Optimization, Quality, Industrial Management, and Science Management (AW)</p> LPPM Universitas Putera Batam en-US Computer and Science Industrial Engineering (COMASIE) 2715-6265 PENERAPAN KLASIFIKASI CITRA PADA IDENTIFIKASI OBJEK DENGAN PAKAIAN SAFETY MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK DI PT JAYATAMA SAFETINDO. https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9647 <p>Construction workers are essential to project execution but face high risks of workplace accidents, often caused by human factors. Advances in artificial intelligence, particularly image processing, provide opportunities to improve the detection of personal protective equipment (PPE), which is currently checked manually and inefficiently. PPE, such as<br />gloves, helmets, and safety shoes, is vital for worker safety but is often neglected due to discomfort. This study uses Convolutional Neural Network (CNN) algorithms to classify images and verify PPE usage at construction sites. CNN processes spatial information through layers for feature extraction, dimension reduction, and classification. A previous<br />study with Faster R-CNN achieved accuracies of 72.83% with TensorFlow and 88.07% with Faster R-CNN. Using a dataset of 200 images, this research, conducted at PT JAYATAMA SAFETINDO, applies Python and TensorFlow to improve PPE detection accuracy. The results aim to support safer workplaces, enhance productivity, and advance AI applications in safety and identification.</p> David Caslan Nababan Sasa Ani Arnomo Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 1 12 10.33884/comasiejournal.v12i2.9647 PERANCANGAN APLIKASI GRAVIDA MENGGUNAKAN METODE AGILE BERBASIS WEBSITE https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9648 <p><em>This research aims to design and develop the Gravida application using the website-based Agile method. This application is designed to help increase efficiency in monitoring daily activities carried out by gravida, such as examination schedules, exercise, diet, and many more that were previously done manually. With this research, gravida activity data can be managed centrally and provide information that is easily accessible to health workers and gravida itself. System testing is carried out using the blackbox testing method to ensure that each feature developed is functioning properly according to user needs. The result of this study show that gravida application is able to provide an efficient and effective solution in managing pregnancy-related information. With this, Gravida application is expected to be a solution for maternity clinics, especially at the Yustina Midwife clinic, in improving work efficiency and service quality to Gravida</em></p> Wafiq Azizah Rahmat Fauzi Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 13 22 10.33884/comasiejournal.v12i2.9648 ANALISIS PERBAIKAN POSTUR KERJA PADA UKM DELTA MEBEL BATU AJI https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9664 <p><em>Delta Furniture UKM is an SME engaged in the furniture product industry in Batam. The purpose of the study is to analyze and identify the work posture of employees using the REBA method and provide suggestions for improvement in the most risky work processes based on REBA assessments. It is important to correct poor posture and working position because poor posture can lead to musculoskeletal disorders (MSDs). The methods used in this study are Nordic Body Map (NBM) and Rapid Entire Body Assessment (REBA). Based on the results of interviews conducted with delta furniture employees, workers complained of pain in the waist and neck because the size of the table was too low so that employees worked in a squatting position and lowered their heads for a long time. In the wood cutting process, the REBA score was obtained 7.3 so that it was in the medium category and needed to be repaired, in the wood boring process the REBA score was obtained 5.6, in the medium category and needed to be repaired. The proposed improvement given is to improve work facilities in the form of desks that are in accordance with the anthropometric size of the worker, namely the dimension of the length of the lateral hand reach of 148.3 cm and the height of the standing elbow 88 cm. After repairing the work facility, the posture of the workers in the wood cutting section gets a REBA score of 2 and for the wood cutting part gets a REBA score of 3, so it is in the small category and no more repairs are needed.</em></p> Fenny Widiya Ningsih Sri Zetli Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 23 32 10.33884/comasiejournal.v12i2.9664 SISTEM DETEKSI POLUSI UDARA BERBASIS INTERNET OF THINGS https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9665 <p><em>Air pollution is one of the most pressing environmental issues, especially in urban and industrial areas such as Batam City, which experiences high pollution levels due to vehicle and industrial activities. This condition significantly impacts public health, including an increased potential for respiratory diseases, asthma, and lung cancer. Therefore, an effective solution is required to monitor air quality in real-time. This research aims to develop an air pollution detection and monitoring system based on the Internet of Things (IoT) that can provide accurate and easily accessible information. The methodology employed in this study includes prototyping and literature review to understand air pollution concepts, IoT, and sensor technologies. The system is designed using MQ-2, MQ-135, and GP2Y1010AU0F sensors to detect harmful gases such as carbon monoxide, carbon dioxide, and particulate matter. The data collected by the sensors is transmitted through an IoT network and displayed on the Telegram application. This application enables users to monitor Track air quality in real-time and get alerts when the air quality changes reaches unhealthy levels. The test results show that the system effectively detects air quality changes. Notifications sent via Telegram align with measurement results displayed on the hardware. This system provides a practical and efficient solution for monitoring air quality in various locations, including industrial areas, urban settings, and public spaces. This research is expected To enhance public awareness of the hazards of air pollution and motivate action to reduce its impact.</em></p> Muhammad Nur Sapi'i Sunarsan Sitohang Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 33 42 10.33884/comasiejournal.v12i2.9665 SISTEM PAKAR IDENTIFIKASI JENIS KULIT DALAM PEMILIHAN SKINCARE MENGGUNAKAN METODE CERTAINTY FACTOR https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9668 <p><em>Expert system for identifying skin types in skincare product selection using the certainty factor (CF) method aims to evaluate the user's skin characteristics and provide appropriate skincare recommendations, considering the different needs of each skin type. The CF method was chosen due to its ability to handle uncertainty in decision-making and provide accurate estimations. This system consists of several stages, starting with the collection of knowledge about skin types through questionnaires and interviews with dermatology experts. The collected data is then translated into rules processed using CF to provide diagnoses with the corresponding level of confidence. System testing shows that the skin type diagnosis provided has high accuracy, making this system a useful tool for helping individuals select the right skincare products. The results of this study indicate that this expert system is effective in assisting users in choosing products suitable for their skin types, reducing the risk of skin problems, and contributing to the improvement of skin health.</em></p> Esther Kristin Sihura Rahmat Fauzi Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 43 50 10.33884/comasiejournal.v12i2.9668 IMPLEMENTASI APLIKASI E-KEHADIRAN BERBASIS FACE RECOGNITION DAN GPS MENGGUNAKAN METODE AGILE DEVELOPMENT https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9675 <p>In an era with advanced technology, technology can help with all needs and facilitate various activities. Attendance in a company is one of the measurement values in employee performance, the attendance used in taking attendance is sometimes still manual and the problem is when the location of the workplace is not in one place. Sometimes attendance data can be manipulated by employees due to lack of monitoring. An attendance application is needed that facilitates digital-based attendance to facilitate attendance and attendance. Building attendance using software development with the agile development method by utilizing face recognition technology and GPS (Global Positioning System) to track employee locations and website-based to facilitate access to various devices. The software that will be used to help build the application is figma as an application design, visual studio code to build website-based applications that support various programming languages. The attendance application is built with the aim that employees can take attendance easily and the company can monitor it easily to reduce fraud and data manipulation by employees due to lack of monitoring from the company.</p> Khoirul Hidayanto Rahmat Fauzi Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 51 60 10.33884/comasiejournal.v12i2.9675 SISTEM MONITORING KARYAWAN PT ATHAYASA MANDIRI MENGGUNAKAN GPS BERBASIS ANDROID https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9680 <p><em>PT Athayasa Mandiri, specializing in heavy equipment and forklift maintenance, faces challenges in real-time monitoring of field technicians due to an ineffective manual system, resulting in delays and reduced service quality. To address this, an Android-based monitoring application is developed using GPS for location tracking and Google Maps API for navigation and route optimization. Built with Flutter for a modern interface and Firebase for reliable data storage, the application adopts the Design Thinking methodology to address user needs, including technicians and management. Key features include real-time location verification, travel history reports, and schedule monitoring, minimizing potential fraud while enhancing task coordination. This system improves transparency, accelerates decision-making, and boosts operational efficiency, reducing costs and increasing customer satisfaction. Additionally, the application is designed to support future IoT integration, providing strategic value and strengthening the company’s competitiveness in the industrial market.</em></p> RIDWAN Ridwan Rahmat Fauzi Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 61 70 10.33884/comasiejournal.v12i2.9680 IMPLEMENTASI ME-FINANCE DENGAN METODE AGILE BERBASIS WEB https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9698 <p><em>The financial management practices in mosques still depend on traditional methods, which makes them susceptible to inaccuracies and inefficiencies. This research seeks to create a web-based financial application utilizing the Agile methodology to improve efficiency, accuracy, and financial reporting. The application is targeted at mosque treasurers, incorporating essential features like transaction logging, financial reporting, and graphical analysis. The research approach consists of data gathering, system design, development, testing, and deployment. To ensure both functionality and user satisfaction, Black Box Testing and User Acceptance Testing (UAT) were performed. The findings indicate that the application successfully automates financial management, minimizing errors and enhancing accessibility</em></p> Della Novita Alfannisa Annurrullah Fajrin Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 71 80 10.33884/comasiejournal.v12i2.9698 IMPLEMENTASI APLIKASI E-COMMERCE BERBASIS ANDROID MENGGUNAKAN METODE WATERFALL https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9705 <p>The development of digital technology, especially the internet, has provided convenience in various activities, including e-commerce-based transactions. This research aims to develop an e-commerce application for Markibong, an online store that offers various souvenirs from Batam, such as food, souvenirs, and clothing.This application is built using the open source Android platform to improve the efficiency of marketing, sales, and transaction processes. The development process follows the Waterfall method in the Software Development Life Cycle (SDLC), including needs analysis to the implementation stage. With this application, it is expected to expand customer access and increase revenue for Markibong stores.</p> Dewa Kurniawan Nusantoro Pastima Simanjuntak Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 81 90 10.33884/comasiejournal.v12i2.9705 DETEKSI KALORI PADA CITRA MAKANAN DENGAN ALGORITMA SINGLE SHOT MULTIBOX DETECTOR https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9714 <p>The increasing prevalence of non-communicable diseases such as obesity and diabetes mellitus has become a major public health concern in Indonesia. Uncontrolled food consumption is one of the primary contributing factors to these issues. Therefore, a system is needed to help individuals monitor their calorie intake more effectively. This study aims to develop a food calorie detection system using the Single Shot Multibox Detector (SSD) method. The model is applied to identify and classify food objects with high accuracy. Calorie estimation is performed based on predefined fixed portion rules. The results indicate that the developed system can recognize various types of food in real-time with optimal performance. The implementation of this system is expected to raise public awareness of healthy eating habits and support efforts to prevent non-communicable diseases in Indonesia.</p> Fifin ayu puspitasari Hotma Pangaribuan Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 91 100 10.33884/comasiejournal.v12i2.9714 PENJADWALAN PREVENTIVE MAINTENANCE MESIN DRY VACUUM PUMP DI PT XYZ BATAM https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9731 <p><em>In the manufacturing industry, ensuring a smooth production process is crucial for maintaining company performance. Machines play a vital role as a production factor that significantly affects the continuity of these processes. At XYZ Batam, frequent machine downtimes disrupt production targets. This study aims to determine the optimal machine maintenance intervals to reduce downtime. The Reliability Centered Maintenance II (RCM II) approach was applied using methods such as Failure Mode and Effect Analysis (FMEA), Mean Time to Failure (MTTF), and Mean Time to Repair (MTTR). The downtime data analyzed covers the period from April 2024 to September 2024. The results indicate that thermal overload repairs on dry vacuum pump no.1 and dry vacuum pump no.2 are the most frequent maintenance activities based on Risk Priority Number (RPN) calculations from the FMEA analysis. Preventive maintenance intervals for machines with the highest downtime rates were determined as follows: dry vacuum pump no.1 requires maintenance every 129,12 hours, and dry vacuum pump no.2 every 151,83 hours.</em></p> Ricky Yakub Sembiring Meliala Elsya Paskaria Loyda Tarigan Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 101 108 10.33884/comasiejournal.v12i2.9731 ANALISIS RISIKO KESELAMATAN DAN KESEHATAN KERJA AKTIVITAS PROSES PRODUKSI DI DEPARTEMEN AUTOLATHE PADA PT XYZ https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9739 <p><em>The Autolathe Department is an organizational unit responsible for producing specific products at PT XYZ using CNC lathes. The tasks in this department include machine operation, machine setup, part rework, and deburring/polishing parts, all of which are exposed to sharp objects, chemicals, and extreme dust. The company has implemented safety measures such as introducing Occupational Health and Safety (OHS) during recruitment and providing various types of Personal Protective Equipment (PPE). However, these control efforts have not been fully effective. This study aims to identify risks in all tasks within the Autolathe department using the Hazard Identification, Risk Assessment, and Determining Control (HIRADC) method and Job Safety Analysis (JSA) method, assess the risk levels, and provide recommendations to the company. The data was collected through documentation, observations, interviews, and discussions with production supervisors who also serve as OHS supervisors. The study found three tasks with high overall risk levels, involving 17 risk variables categorized as extreme or high, such as flying debris, cuts from cutter blades, material pinching, and exposure to chemicals and dust. The study recommends improving the system, including strict sanctions for operators who fail to wear PPE according to Standard Operating Procedures and providing additional PPE.</em></p> Reychandro Eryvan Pardomuan Sijabat Welly Sugianto Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 109 118 10.33884/comasiejournal.v12i2.9739 ANALISIS LINE BALANCING PADA PRODUKSI SETRIKA DI PT PHILIPS INDUSTRIES BATAM https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9749 <p><em>This study discusses line balancing in the production of irons at PT Philips Industries Batam, aiming to improve efficiency by addressing bottlenecks in the manufacturing process. The main focus is to optimize task distribution across workstations to minimize non-value-added time and enhance production line performance. Before the improvements, line efficiency was recorded at 94%, but there was still considerable idle time. Data collection was conducted through direct observation and time measurement of 18 operations using the Ranked Positional Weight (RPW) method. The analysis revealed that the implementation of Unique Device Identifier (UDI) labels led to increased operational times, particularly in operations 11, 15, and 18, causing workflow delays. After applying line balancing strategies, efficiency significantly improved from 67.97% to 96.23%. This study demonstrates that better task distribution can reduce production delays and support the achievement of operational targets. Key improvements include a reduction in balance delay and a smoother production flow, resulting in increased overall productivity. These findings highlight the importance of line balancing in optimizing manufacturing performance and provide valuable insights for industries seeking to enhance competitiveness through a more efficient production system.</em></p> Dian Pratiwi Elsya Paskaria Loyda Tarigan Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 119 128 10.33884/comasiejournal.v12i2.9749 RANCANG ULANG MESIN PENCETAK BATAKO PADA PT BODEM MAS JAYA https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9770 <p>PT Bodem Mas Jaya is a company engaged in concrete block production. However, the company faces challenges in meeting production targets due to the limitations of its block-making machine, which can only produce two blocks per process. This study aims to redesign the block-making machine to improve efficiency and production capacity. Kansei Engineering was applied to identify user needs, resulting in 16 Kansei words such as ergonomic, productive, precise, and modern. The redesign features a four-mold system, enabling the machine to produce four blocks simultaneously, doubling production capacity. Additional features such as rubber barriers, tray holders, and mold lifters enhance stability, minimize material waste, and maintain block quality. The new design also emphasizes ease of maintenance, safety, and operational efficiency. With this redesign, the company is expected to meet the increasing market demand effectively.</p> Joko Saputro Ganda Sirait Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 129 138 10.33884/comasiejournal.v12i2.9770 ANALISIS PERBANDINGAN KINERJA ALGORITMA MACHINE LEARNING BERBASIS FEATURE SELECTION DALAM DETEKSI SERANGAN BOTNET https://forum.upbatam.ac.id/index.php/comasiejournal/article/view/9778 <p><em>Internet has experienced significant development. Increasing devices connected to internet makes security against cyber attacks a critical issue, thus creates opportunities for cyber attackers, one form of those attack is botnets. In Indonesia, Botnets is the highest traffic anomalies in 2022 by BSSN. High number of attacks because detecting botnet can be challenging, difficulty of detecting attacks and low level of detection accuracy means that normal data sometimes considered an attack, so choosing method that can handle this is very important. Machine learning algorithms are able to study network data traffic and identify suspicious activity, this makes machine learning an effective method. Machine learning based on feature selection has an accuracy of above 90% in detecting DDoS attacks on datasets and machine learning algorithms are also able to detect attack data and normal data. Thus, in this research machine learning algorithms such as K-Nearest Neighbors, Support Vector Machine and Naive Bayes will be applied to dataset containing botnet and normal data to explore how machine learning algorithms can effectively detect botnet attack patterns and normal data. This research compares the performance of commonly used machine learning algorithms to find which one effective for detecting botnet attacks in existing datasets.</em></p> Rio Khoo Koko Handoko Copyright (c) 2025 Computer and Science Industrial Engineering (COMASIE) 2025-02-03 2025-02-03 12 2 139 148 10.33884/comasiejournal.v12i2.9778