<![CDATA[This study presents the development of a web-based industrial waste prediction system using the Autoregressive Integrated Moving Average (ARIMA) method to forecast the volume of liquid and solid waste generated by PT Pupuk Iskandar Muda (PIM). The predictive model is built upon historical waste data collected between 2020 and 2023, serving as the foundation for the statistical analysis. The system is developed using the Flask web framework, offering an interactive and user-friendly interface, while SQLite3 is employed as a lightweight local database solution for efficient data handling. The ARIMA (1,1,1) model was selected based on stationarity testing and examining ACF and PACF patterns. The results suggest that the model can moderately capture prediction trends, although limitations in accuracy are evident. For 2024, liquid waste is projected to decrease from 30,600 tons in January to 29,400 tons in December. In contrast, solid waste displays a more stable trend, with an average monthly generation of approximately 23.2 tons. Model performance was evaluated using the Mean Absolute Percentage Error (MAPE) method, yielding high error rates—166.11% for liquid waste and 100% for solid waste, highlighting the significant impact of data quality and completeness on prediction accuracy. The system generates visual outputs through interactive graphs and tables accessible via a web browser, supporting data-driven decision-making. This research is a predictive tool for PT PIM and a reference for future development of technology-driven waste management systems to promote environmental sustainability.]]>
Copyrights © 2025
