The prediction of stock management for Farmasi Chendering

In the retail pharmacy sector, data analytics is crucial for enhancing inventory control and operational efficiency. This project focuses on Farmasi Chendering, aiming to develop a predictive stock management system using ABC-VEN analysis and the J48 decision tree algorithm. The adapted CRISP-DM methodology guided the development process, encompassing data preparation, modeling, evaluation, and deployment. Historical sales data comprising 2 years of transactions were analyzed in two experimental setups: one without ABC-VEN classification and one with it. After balancing the dataset using SMOTE, the ABC-VEN-enhanced model achieved an improved prediction accuracy of 79.11%, compared to 73.41% without ABC-VEN. This confirmed the model's effectiveness in refining stock forecasts. An interactive dashboard was developed to visualize 2023 sales data, inventory classification, and 2024 stock predictions, enabling informed and timely restocking decisions. Expert evaluations involving pharmacy professionals highlighted the dashboard's practicality, with overall positive feedback and suggestions for minor enhancements, such as more detailed summaries.In conclusion, this project demonstrates how integrating predictive analytics and inventory classification can improve stock management in community pharmacies. Future improvements may include real-time data integration, expansion to mobile platforms, and broader product coverage to support sustainable decisionmaking.