Journal of Information Systems and Informatics (Journal-ISI)
Not a member yet
580 research outputs found
Sort by
Traffic Violation Clustering Using K-Medoids and Word Cloud Visualization
Full text linkTraffic is the space for people to move around, including both drivers and pedestrians. According to data from the Central Statistics Agency in 2020, the number of motor vehicles in Makassar City was recorded by type: 248,682 passenger cars, 17,501 buses, 85,968 trucks, and 1,338,306 motorcycles, with a tendency for an increase in the following year. The high number of vehicle users can certainly affect the rising traffic violation rates on the road. This study aims to classify traffic violation types in Makassar City by utilizing the K-Medoids algorithm and to visualize the clustering results using Word Cloud, which is expected to provide information related to patterns of traffic violation clusters. This study uses a case study from the Traffic Police Department of Makassar City in 2021, with a total of 5,893 traffic violation cases. The data used is ticket data consisting of article and vehicle type features. The clustering results show that motorcycles and minibuses are the most frequently involved in traffic violations. Motorcycles (R2) are not only dominated by violations related to the use of standard SNI helmets but also significantly involved in violations related to incomplete requirements and the possession of SIM/STNK (Driver's License/Vehicle Registration) and failing to meet roadworthiness standards such as mirrors, headlights, horns, etc. Passenger vehicles, especially minibuses and cars, also dominate traffic violations. The violations involve not only the use of seat belts for R4 vehicles but also violations such as not having complete STNK, not being able to show SIM, failing to display the Vehicle Registration Mark (TKB), and others. The results of this study demonstrate that the clustering obtained is very strong, as evidenced by the high Silhouette Score of 0.867 at k = 9
Strategic IS/IT Planning for Enhanced Competitiveness and Operational Efficiency at PT. Songgo Jati Baru: Applying the Ward and Peppard Method
Full text linkThis study designs an integrated IS/IT strategy to enhance PT. Songgo Jati Baru's operational efficiency and global competitiveness in the trading and services sector. A qualitative approach was employed, utilizing data collection methods such as focus group discussions, document analysis, observation, and interviews. Analysis was conducted using the robust Ward and Peppard method, which incorporates SWOT analysis, Gap analysis, and the McFarlan Strategic Grid. The findings revealed that the company faced significant challenges, including a lack of system integration, limited data analytics capabilities, and suboptimal digital marketing strategies. To address these, the research recommends a cloud-based Enterprise Resource Planning (ERP) system for comprehensive business process integration, a Vendor Management System (VMS) for efficient collaboration, and a Customer Relationship Management (CRM) system for data-driven marketing, with a phased implementation planned for 2026-2028. This comprehensive strategy, underpinned by robust cloud infrastructure and continuous staff training, is poised to not only significantly enhance PT. Songgo Jati Baru's operational efficiency and global market reach but also to solidify its competitive position and ensure sustainable growth in the dynamic trade and services sector
A Novel UX-Centered ITSM Framework for Technology Startups: Beyond Traditional Service Management
Full text linkThis research explores the integration of User Experience (UX) principles into IT Service Management (ITSM) frameworks within resource-constrained B2B SaaS technology startups. Through a comprehensive qualitative case study methodology involving semi-structured interviews with seven stakeholders, participatory observation across 12 sessions, and systematic document analysis at a Jakarta-based startup serving SMEs, we uncovered a critical paradox: companies selling superior UX solutions to clients often neglect these principles in internal IT management. The primary contribution is a novel adaptive UX-Centered ITSM conceptual model featuring three interconnected layers: Core Principles, Implementation Domains, and Operational Elements, designed for incremental implementation based on startup capacity. Unlike rigid existing ITSM frameworks, this model introduces a prioritized approach with "Must Have," "Should Have," and "Can Be Added" categorizations specifically tailored for startup contexts. The research identified five contextual factors influencing implementation: organizational culture, leadership structure, resource limitations, team dynamics, and SME client characteristics. Findings reveal that UX-centered ITSM not only addresses internal operational challenges but creates strategic alignment between internal practices and external value propositions, forming the foundation for market credibility and business sustainability. This framework provides startup managers and IT practitioners with an actionable roadmap for transforming ad-hoc internal systems into user-centered services that support operational excellence while enhancing competitive positioning in digital transformation markets
Examining ICT Interventions for Rural Health System Connectivity: Challenges and Gaps for Improvement: A Systematic Review
Full text linkCommunity healthcare interventions in Low- and Lower-Middle-Income Countries (LLMICs) frequently face record management issues that hinder effective linkage between community services and national health systems, contributing to persistently high mortality rates. This study aimed to identify and analyze ICT-based community health interventions implemented in LLMICs, evaluate their effectiveness, and explore challenges limiting their impact. A comprehensive literature search was conducted in June 2024 across ACM Library, PubMed, ScienceDirect, and Google Scholar, focusing on studies published between January 2019 and May 2024. Inclusion criteria targeted ICT-based interventions conducted in LLMICs, available in English, with accessible full texts and clearly defined ICT components. Of the 792 records initially screened, only 9 met the eligibility requirements. Most interventions addressed individual components such as data collection, monitoring, consultations, referrals, and reminders. However, they often lacked integrated systems for data management, continuity of care, and follow-up, limiting their long-term effectiveness. While the review was restricted to open-access studies, the findings offer crucial insights into the design and implementation of ICT-based health solutions. The absence of process integration in current interventions remains a major barrier. Future research and policy development should focus on designing comprehensive, integrated ICT frameworks to strengthen community-to-health system linkages and improve health outcomes in LLMICs
Development and Capability Evaluation of a Firebase-Based Pharmacy Inventory System Using COBIT 2019
Full text linkPharmacy inventory control systems require proper and timely data management to be able to supply medicines and function as such. The study conceptualized and tested a Firebase-based system of pharmacy inventory control. The system was conceptualized in the web and Android platforms with the key objectives of enabling real-time synchronization, tracking, and reporting functionalities. Capability measurement from utilizing the COBIT 2019 method was utilized in evaluating system governance and operational performance. Four key processes are BAI03 (Identification and Build of Managed Solutions), DSS01 (Managed Operations), DSS02 (Managed Service Requests and Managed Incidents), and MEA01 (Managed Performance Monitoring) were chosen shortlisted and mapped to system indicators. Seven pharmacy employees were subjected to the assessment with a Likert-scale questionnaire. The results showed that three processes attained Capability Level 4 (Predictable) and one attained Capability Level 5 (Optimizing), i.e., the system performs predictably and allows continuous improvement. Weakness points despite deployment of the system with proof of handling data with ease and responsiveness were the fact that the sample size of respondents was small and one pharmacy only had it deployed. Much more must be done to experiment with the system in different environments and explore integration with third-party platforms for further scalability and adherence to governance
Comparison of RNN and LSTM Classifiers for Sentiment Analysis of Airline Tweets
Full text linkThis study examines the application of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) models for sentiment analysis of airline-related tweets, focusing on customer feedback directed at U.S. airlines on the X platform (formerly Twitter). The objective was to utilize these deep learning models to identify sentiment trends within text data and compare their performance in terms of computation time. The analysis was conducted on a 14,640-imbalanced dataset of classified tweets from February 2015 as positive, negative, or neutral. Both models were trained under identical conditions using Term Frequency-Inverse Document Frequency (TF-IDF) and Word2Vec for feature extraction. LSTM achieved 74% accuracy with AUC scores of 0.84, 0.90, and 0.89. RNN achieved 72% accuracy with AUC scores of 0.78, 0.87, and 0.85. In terms of time efficiency, RNN outperformed LSTM, requiring 57.16 seconds for training and 0.52 seconds for testing, compared to LSTM’s 82.40 and 0.82 seconds. Time performance was also evaluated per sentiment class, and RNN consistently outperformed LSTM. These results highlight the trade-off between accuracy and computational cost. Limitations include dataset imbalance and LSTM’s slower computation due to its internal gate mechanisms. Future work could prioritize integrating hybrid models and may use data imbalance techniques to improve sentiment classification
Deep Learning and Statistical Models to Analyse Online Misinformation and Hate Speech Impact on African Youth
Full text linkThis study examines the perceptions, behaviour, and digital experiences of African youth in relation to online misinformation and hate speech. Using a large-scale, cross-national survey with 10,005 valid responses, the research relies on both statistical clustering and deep learning-based autoencoder models to group youth together based on their trust in information, concern about misinformation, verification behaviours and platform usage. The dual-method analysis highlights three distinct behavioural and attitudinal clusters of youth, denoting different levels of digital skeptical engagement, exposure, and civic engagement. The findings highlight the heterogeneity within the youth population and emphasize that a one-size-fits-all approach to combating misinformation is insufficient. Notably, youth with high concern also demonstrated strong verification habits, while less engaged clusters exhibited low concern and limited digital resilience. These insights offer a foundation for designing cluster-specific interventions and media literacy strategies that are regionally and behaviourally responsive. This combination advances research through unsupervised deep learning on large social survey data, as well as demonstrating the utility of deep learning in revealing latent behaviours. The implications of this study's findings are timely for educators, policy makers and digital platforms more broadly, that want to foster informed and safe digital participation for African youth. As scalable, data-driven framework is a contribution towards an inclusive digital policy package for varied youth realities that exist in an African context
Expert System for The Diagnosis of Depression in Students Using Certainty Factor Method: A Case Study of Ngudi Waluyo University
Full text linkDepression is a growing mental health concern among university students, often fueled by academic pressure, social demands, and personal stress. This study presents the development of an expert system using the Certainty Factor (CF) method to diagnose depression specifically among students at Ngudi Waluyo University. The system categorizes depression into mild, moderate, and severe levels based on 12 validated symptom statements and expert-defined diagnostic rules. Implemented with PHP, JavaScript, and CSS, the system offers a user-friendly, accessible, and anonymous platform for self-assessment. Testing yielded an accuracy rate of up to 79% in diagnosing depression severity and a 71.7% user satisfaction rate based on a User Acceptance Test (UAT) involving 32 students. Results demonstrate that the system can effectively support early detection and mental health awareness within academic environments. Despite some limitations in UI and feedback depth, the expert system shows strong potential for broader application and further enhancement
Hybrid Unsupervised Machine Learning for Insurance Fraud Detection: PCA-XGBoost-LOF and Isolation Forest
Full text linkInsurance fraud poses a significant threat to the financial stability of insurance companies, resulting in substantial economic losses. To combat this issue, this study proposes a novel unsupervised machine learning hybrid algorithm, integrating Principal Component Analysis (PCA), Extreme Gradient Boosting (XGBoost), Local Outlier Factor (LOF), and Isolation Forest. This hybrid approach aims to improve the detection accuracy of insurance fraud by combining the strengths of each individual algorithm. Experimental results a real-world insurance dataset demonstrate a detection accuracy of 92%, precision of 92% and recall of 96%. Our experimental results demonstrate that the proposed hybrid algorithm outperforms existing state-of-the-art methods, achieving a higher detection rate and reducing false positives. This research contributes to the development of effective insurance fraud detection systems, ultimately helping insurance companies to minimize financial losses and improve their overall profitability
Digitalization of Tobacco Taxation in Bangladesh: Reducing Evasion and Enhancing Public Health
Full text linkThis study investigates the worldwide practice of digital tax systems to develop a digital taxation model for Bangladesh and identifies the prospects and challenges of implementing this model to address tax evasion. This study applies a qualitative approach to research. 30 in-depth interviews (IDI) with retailers of tobacco products and 10 key informant interviews (KII) with National Board of Revenue (NBR) officials, Tobacco control activists, and tax experts have been conducted using semi-structured interview guidelines. Secondary data has been collected from various reports, and journal articles. Data has been analyzed using the thematic analysis technique. The potential benefits of implementing a digitalized tax system encompass minimization of tax evasion, better monitoring, tracking, and tracing systems, transparent tax administration, and developing an efficient tax collection system. Despite the many advantages of digital tax systems, several challenges must be addressed. These include administrative resistance due to a lack of skilled manpower and modern infrastructure and the difficulties associated with registering tobacco companies and farmers. The implementation of proposed digital taxation model is expected to control tobacco tax evasion which ultimately increases prices and contributes to the overall goal of reducing tobacco consumption and enhance public health in Bangladesh