IJCONSIST JOURNALS

Performance Evaluation of XGBoost and Random Forest Models in Visibility Prediction at Juanda Airport

Ananda Amelia Pramaisita — 2026-03-01

Predicting meteorological visibility is critical in enabling transportation safety and weather watch systems. The current study compares the accuracy of the Random Forest and XGBoost algorithms in performing time series prediction for visibility based on BMKG Juanda, Sidoarjo hourly observation records for one year. Process analysis only uses visibility as a primary variable. Preprocessing of the data involved handling missing values, normalization, and dividing the data into training and test datasets. Model training and hyperparameter tuning were followed by model evaluation using a combination of MAE, RMSE, and MAPE indices. From the results, it is found that Random Forest had an MAE of 808.54, RMSE of 1,312.64, MAPE of 21.09%, and a computation time of 1.02 seconds, while XGBoost had an MAE of 808.81, RMSE of 1,323.12, MAPE of 21.47%, and a computation time of 1.36 seconds. As such, Random Forest is proposed as a more efficient model for predicting visibility at Surabaya's Juanda Airport; however, XGBoost remains a consideration for applicability when there is excessive variability in the data.

Development of Design Digital Application System for Laundry Partner Services: A Smart Village Initiative

Nicholas Napitupulu — 2026-03-01

The development of smart village concepts promotes the use of digital technology to enhance laundry service efficiency at the village level. This paper outlines the development of a digital application system addressing manual challenges such as stock recording, payments, and garment tracking at partner laundry managed by Virna Indri Luthfia. The system supports customer management, orders, payments, and real-time reporting via mobile features for customers, partners, and admins.

Event-Driven Conversational Workflow Architecture for Academic Service Automation Using n8n

Bonda Sisephaputra — 2026-03-01

The digital transformation of higher education institutions has increased the demand for efficient academic service management systems. However, many existing digital academic service platforms primarily focus on providing online interfaces while the underlying administrative processes remain fragmented and manually coordinated. As a result, academic service requests such as document submissions and information inquiries often lack process transparency and structured workflow management.

This study proposes an event-driven conversational workflow architecture for automating academic service processes. The proposed architecture integrates a conversational interface through WhatsApp messaging services, a workflow orchestration engine implemented using the n8n automation platform, and a cloud-based database using Supabase PostgreSQL. The system consists of three main workflows: knowledge base generation, information services using a Retrieval-Augmented Generation (RAG) approach, and document request services with multi-level approval processes.

The research adopts the Design Science Research methodology to design and evaluate the proposed system architecture. A prototype system was implemented to demonstrate how academic service requests can be initiated through conversational interfaces and processed through event-driven workflow automation. The evaluation results show that the proposed architecture enables automated workflow orchestration, structured service state management, and improved service transparency.

Design and Development of a GAS-Based E-Commerce Application with UI/UX Optimization Using a User-Centered Design Approach

Wahyu Nofiyan Hadi — 2026-03-01

Micro, Small, and Medium Enterprises (MSMEs) offering custom embroidery and tailoring services face significant challenges with digitizing orders. Conventional e-commerce platforms primarily serve ready-made products. They lack flexibility for custom workflow options, like letting customers submit their own designs or bring in clothing for embroidery. MSMEs are often burdened by ongoing server maintenance costs. This study seeks to design and implement a web-based e-commerce information system using a User-Centered Design (UCD) approach. The case study is applied to Cindy Bordir in Probolinggo. To address the research gap regarding high-cost and rigid platforms, the application introduces a zero-cost serverless architecture. This uses Google Apps Script (GAS) and Google Spreadsheet as the database. A "Hybrid Checkout" flow bridges the digital catalog experience with localized negotiation habits via WhatsApp. Results show that applying UCD via a mobile-first interface, chip-style product variations, and micro-interaction feedback effectively support custom transaction needs. Functional testing confirmed a 100% success rate for all features. Usability evaluations using the System Usability Scale (SUS) achieved an "Excellent" score of 84.5, indicating high user satisfaction. The serverless architecture paired with the hybrid checkout approach proved to be a highly cost-effective and manageable solution for MSMEs. It eliminates traditional hosting fees while maintaining operational reliability.

The Implementation of a Web-Based E-Commerce Information System for Gold Trading at Bintang Emas Store

Moh. Fadel — 2026-03-01

The development of information technology has encouraged digital transformation in various business sectors, including gold trading businesses. Bintang Emas Kraksaan Store still applies conventional transaction processes, causing limitations in product information delivery, transaction recording, and gold price updates. This study aims to implement a web-based e-commerce information system for gold trading to improve transaction efficiency and digital data management. The system was developed using the Waterfall method, consisting of requirement analysis, system design, implementation, testing, and maintenance stages. The developed system provides several features such as product management, gold price updates, online transactions, user authentication, and sales reporting. System testing was conducted using Black Box Testing and User Acceptance Testing. The results indicate that the system successfully improves transaction efficiency, facilitates digital transaction management, and provides real time gold price information. Based on the User Acceptance Testing (UAT)(UAT)results, the system is considered acceptable and easy to use by users

Enhancing Village Decision-Making through an AHP-Based Decision Support System with Centralized Database Integration

Anisa Nurul Wilda — 2026-03-01

Digital transformation at the local government level requires reliable human resources, particularly in terms of digital literacy. However, the evaluation of digital literacy is often conducted manually and subjectively, making the results less consistent and difficult to use as a basis for data driven decision making. This study aims to develop a decision support system by integrating the Analytical Hierarchy Process method with a centralized database to support village decision making. The case study was conducted in Clarak Village, Leces District. The proposed system is designed to evaluate digital literacy based on several criteria, including the ability to use digital technology, understanding of digital information, digital security and ethics, and the use of technology in work activities. The research method applies a Hybrid Waterfall SPK approach consisting of requirement analysis, system and database design, implementation, AHP integration, functional testing, validation, and usability evaluation. The centralized database is used to manage user data, criteria, weights, and evaluation results in an integrated manner. The expected result of this research is a prototype of an AHP based decision support system that can produce objective, measurable, and consistent digital literacy evaluations to support effective village level decision making.

Analyzing Public Sentiment Toward Probolinggo UMKM Products on TikTok Through Naïve Bayes Classification and Data Visualization

Umi Diantika — 2026-03-01

The development of social media, particularly TikTok, has become one of the main platforms for the public to provide opinions and reviews on MSME products. This research aims to analyze public sentiment towards MSME products in Probolinggo based on user comments on TikTok using the Naïve Bayes Classification method. Research data was obtained through scraping TikTok comments, followed by preprocessing steps including cleaning text, case folding, tokenizing, stopword removal, and stemming. The data was then labeled as positive, neutral, and negative sentiment before feature extraction using TF-IDF. The classification process was carried out using the Naïve Bayes algorithm combined with SMOTE on training data to overcome class imbalance, and evaluated using pure test data. The results showed that the Naïve Bayes method was able to classify sentiment with a final accuracy rate of 66.67%. The analysis showed that neutral sentiment dominated public opinion with 365 comments (47.04%), followed by positive sentiment with 341 comments (43.94%) and negative sentiment with 70 comments (9.02%). Data visualization indicated that most users responded very positively to MSME Probolinggo products on TikTok. This research is expected to help MSME actors understand consumer perceptions and improve digital marketing strategies based on social media