Perbandingan Metode Single Exponential Smoothing dan Double Exponential Smoothing untuk Peramalan Omzet Penjualan Café

Abstract

The culinary industry of Micro, Small, and Medium Enterprises in Bali faces challenges in operational decision-making, which still tends to rely on intuition due to the suboptimal utilization of historical sales data. This study aims to analyze daily revenue patterns, apply the Single Exponential Smoothing method to forecast future revenue, and develop an interactive dashboard as a data visualization tool. As a comparison to obtain the best accuracy, the SES method is evaluated against the Double Exponential Smoothing method. This research adopts a quantitative approach using a case study of Ruby’s Coffee and Kitchen in Bali, based on daily revenue data from March 2025 to February 2026. The analysis results using the Walk-Forward Validation approach indicate that the SES method with a parameter α = 0.1 achieves the best accuracy with the lowest average Mean Absolute Percentage Error of 25.46%, outperforming the DES method, which produces an average MAPE of 30.73% at the parameter α = 0.1. In addition, usability testing results yield an average score of 4.67 out of 5, suggesting that the dashboard provides an excellent level of ease of use in supporting operational monitoring of the café. The contribution of this study lies in the integration of forecasting methods with a Streamlit-based interactive dashboard, which presents predictive results in a visual and informative manner, thereby assisting management in more efficient operational planning and supporting more accurate data-driven decision-making.