Rancang Bangun Aplikasi Web Pencatatan Transaksi Dan Stok Barang Pada Warung Kelontong Dengan Prediksi Kebutuhan Stok Menggunakan Machine Learning

Traditional convenience stores are a form of micro-enterprise that often rely on manual transaction and inventory recording or, in some cases, no recording at all. This condition frequently leads to problems such as stock discrepancies, transaction calculation errors, and the absence of structured sales reports. This study aims to design and develop a web based application to assist convenience store owners in recording transactions, managing inventory, and generating integrated sales reports. The system development method used is the Software Development Life Cycle (SDLC) with the Waterfall model, which consists of requirements analysis, system design, implementation, testing, and maintenance stages. The developed application is equipped with a digital cashier feature, item and supplier management, goods-in recording, low-stock notifications, and receipt printing via a Bluetooth printer. In addition, the system incorporates a stock demand prediction feature using the AutoRegressive Integrated Moving Average (ARIMA) algorithm based on historical transaction data. The prediction model was evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE), and the results indicate that the ARIMA model demonstrates stable and consistent performance, making it suitable for supporting restock decision making. Functional testing using the black box method confirms that all system features operate according to the intended usage scenarios, while the System Usability Scale (SUS) evaluation achieved a score of 71.56, which is classified as good, indicating that the system is acceptable and easy to use by users.