<![CDATA[Fluctuations in shallot prices in Indonesia create uncertainty within the agricultural supply chain and affect farmers, traders, and policymakers. This condition highlights the need for analytical mechanisms capable of accurately monitoring and predicting price dynamics. This study develops a web-based shallot price prediction system using the Rapid Application Development (RAD) method, with the best-performing model obtained from the training process being a combination of Long Short-Term Memory (LSTM) and CatBoost. The model is designed to process historical data along with non-sequential variables including price, production, rainfall, inflation, the Consumer Price Index (CPI), and seasonal indicators using a five-year dataset compiled from various official government sources. The trained model is integrated into a Flask-based backend to generate the next 7-day price forecasts. The system allows users to upload datasets, execute prediction processes, and analyze outputs through interactive charts and prediction tables. The evaluation shows that the model achieves strong performance, indicated by a MAPE of 6.71% and an RMSE of 0.029120, reflecting good accuracy and alignment with the seasonal patterns of shallot prices. Black-box testing confirms that all system functions operate as expected. The RAD method contributes to accelerating the development process through continuous iteration, resulting in a lightweight, responsive, and user-friendly system for non-technical users. Consequently, this system has the potential to serve as a decision-support tool for monitoring and anticipating shallot price dynamics at both regional and national levels.]]>
