Jurnal Sistem Informasi (JUSIN)

PENGELOMPOKAN PENJUALAN PRODUK DENGAN MENGGUNAKAN K-MEANS CLUSTERING : STUDI KASUS ANALISIS PENJUALAN COFFEE SHOP OLEH KAGGLE.COM

Sifa Rismawati Sifa — 2025-06-25

This study aims to group sales products in a coffee shop based on transaction data using the K-Means Clustering algorithm. The dataset from Kaggle.com includes the attributes product_id, transaction_qty, and unit_price. This method was chosen because of its ability to identify sales patterns in grouping products into three main clusters including high, medium, and low sales. The research process includes data collection, pre-processing, normalization, determining the optimal number of clusters, to evaluating the results using a Silhouette Score of 0.65. These results indicate that the K-Means method is effective in providing product segmentation that can be used as a basis for making business decisions, in optimizing stock and data-based marketing strategies.

OPTIMASI ALGORITMA RANDOM FOREST UNTUK MENINGKATKAN AKURASI PREDIKSI INDEKS MASSA TUBUH (BMI)

Eva Juliani — 2025-06-25

Accurate Body Mass Index (BMI) prediction is essential for detecting obesity risks and related diseases. This study optimizes the Random Forest algorithm to enhance BMI prediction accuracy through hyperparameter tuning and feature selection. The dataset used is Obesity: Raw and Synthetic Data, which includes demographic and lifestyle variables. After undergoing subsetting, label encoding, and data imbalance handling using SMOTE, the model was trained using Random Forest and evaluated with accuracy, precision, recall, and F1-score metrics. The results indicate that the optimized model achieved 90% accuracy, with precision and recall of 0.89. Additionally, the feature importance analysis identified weight, height, and dietary habits as the most influential factors in BMI prediction. These findings confirm that optimizing the algorithm enhances model reliability in BMI classification and can be applied in data-driven health monitoring systems. This research is expected to contribute to the development of digital health applications and more accurate early obesity detection systems.

OPTIMALISASI SISTEM INFORMASI BERBASIS AGILE UNTUK PENGELOLAAN BIAYA VARIABEL DAN EXCHANGE RATE DI INDUSTRI FMCG

Muhammad Rizky Saputra — 2025-06-25

In the Fast Moving Consumer Goods (FMCG) industry, managing variable costs and exchange rates is a major challenge that affects operational efficiency and business decision-making. This study aims to develop an Agile-based information system that enables real-time data integration to improve transparency and accuracy in recording variable costs and exchange rates. The system was developed using the Agile Software Development methodology, consisting of Product Backlog, Sprint Planning, Daily Scrum, Development, Testing, and User Acceptance Testing (UAT). Key features include a variable cost and exchange rate dashboard, which provides structured data presentation, and an input form that facilitates users in entering and managing data according to operational needs. Testing was conducted through UAT involving the System Developer, IT Product team, and production team users. The results indicate that the system enhances cost recording efficiency, provides better visibility of exchange rate fluctuations, and accelerates data validation processes.

PENGEMBANGAN APLIKASI BERBASIS WEB UNTUK PERAWATAN MOBIL DAN PENGELOLAAN BENGKEL DENGAN METODE PROTOTIPE

Aisha Gemala Jondya — 2025-06-25

This study presents the development of CekMobilmu.com, a web-based application designed to support car owners and automotive workshop in Jakarta by facilitating the digital recording of vehicle maintenance and repair activities. In response to the increasing technological advancement and rising demand for efficient service documentation, the application provides features for logging service history, monitoring vehicle condition, and managing workshop operations. Employing the Prototype Method, the research was conducted through sequential stages of data collection, system design, implementation, and evaluation. Data were obtained using questionnaires, interviews, and comparative analysis with existing similar applications. The evaluation phase demonstrated that the application effectively assists users in tracking maintenance schedules and enables workshops to record and manage repair tasks more efficiently. CekMobilmu.com addresses the communication gap between vehicle owners and service providers, offering a centralized and practical solution for vehicle maintenance management.

ANALISIS FAKTOR SOSIAL EKONOMI YANG MEMPENGARUHI RENDAHNYA CAPAIAN PENDIDIKAN DI INDONESIA MENGGUNAKAN KOMBINASI METODE DATA MINING

Diana Yusuf — 2025-06-25

Educational inequality remains a persistent issue in Indonesia, particularly in regions with challenging socio-economic conditions. This study aims to analyze how various socio-economic factors influence the average years of schooling across Indonesian provinces using a combination of K-Means Clustering and Decision Tree algorithms. The dataset includes indicators such as poverty rate, gross regional domestic product (GRDP), per capita expenditure, and life expectancy, obtained from official national statistics.

K-Means Clustering was employed to segment provinces into three distinct groups based on socio-economic similarities. The clustering revealed clear disparities among regions, where the most disadvantaged cluster showed significantly lower education levels. Subsequently, the Decision Tree algorithm was used to classify the average years of schooling, identifying per capita expenditure, life expectancy, and socio-economic cluster as the most influential variables.

The combined approach allows for both segmentation and interpretation, providing data-driven insights that are accessible and actionable for policymakers. The findings highlight the importance of targeting socio-economic improvements as a strategy to enhance educational outcomes. Ultimately, this study underscores the value of integrating unsupervised and supervised machine learning models in addressing complex social issues in education.

PENGARUH TIKTOK ADS DAN BLACK CAMPAIGN TERHADAP PERILAKU PEMBELIAN KONSUMEN DI KOTA BATAM

R. Ilham Ramizan Fahris — 2025-06-25

The use of social media platforms such as TikTok as a tool for digital marketing continues to grow along with the increasing number of users, both for building brand awareness through TikTok Ads and indirectly through black campaigns. This study aims to analyze the influence of TikTok Ads and Black Campaigns on consumer purchasing behavior in Batam City. A quantitative method was employed by distributing online questionnaires to active TikTok users using stratified disproportional random sampling. Data analysis was conducted using SPSS software, including validity and reliability tests, multiple linear regression, and classical assumption testing. The results show that both TikTok Ads and Black Campaigns have a positive and significant influence on consumer purchasing behavior, with B coefficients of 0.360 and 0.390, respectively. These findings indicate that although black campaigns are negative in nature, their exposure can trigger consumer curiosity about the highlighted product. The implications of this study suggest that business actors need to be more selective and adaptive in developing digital marketing strategies, as consumer perceptions are influenced not only by positive promotions but also by the negative narratives circulating online. Therefore, ethical and fact-based communication strategies are key to maintaining consumer trust and brand reputation in the digital era.