Autoregressive Integrated Moving Average (ARIMA) Application with Dashboard Visualization to Predict Retail Sales Data

Many business actors have used information technology to make business processes easier. Tambah Ilmu Store in Sidoarjo sells stationery and office supplies, and they use a cashier application to record the items they have sold. When a business runs a process to meet customer needs, there are unavoidable obstacles, such as empty product stock due to high customer buying interest and excessive product stock purchases, which cause stock to pile up. Therefore, data mining is needed to manage the history of product sales that were previously only stored, because this can help stores see sales patterns and predict the actual amount of product stock. This study uses a time series prediction algorithm, namely ARIMA (Auto Regressive Integrated Moving Average), to predict products that will be sold based on sales history from January 2022 to February 2024. In this study, a data mining framework known as the Cross Industry Standard Process for Data Mining (CRISP-DM) is used. This research is quantitative, using secondary data. The results of the study showed that of the five main products, only one showed good prediction results: item krts182 had a MAPE value of 16.40%, while the other products showed fairly good prediction results: item code amp003 had a MAPE value of 23.55%, bktls0003 43.98%, krts281 33.48%, and blpn021 30.96%. It is expected that the existence of sales predictions will help the store in choosing stock to buy according to the sales pattern and predictions made. Using Tableau tools, the results of the prediction analysis are visualized into diagrams and graphs. It is hoped that the presence of sales predictions will help the store in choosing stock to buy according to the sales pattern and predictions made.