Machine Learning Approach to Analyzing and Predicting Revenue Trend in Mineral Water Sales

Mineral water sales in Indonesia have experienced rapid annual growth, driven by factors such as population increase, lifestyle changes, and a 40.64% preference for bottled water among households. The increasing demand and sales of mineral water present new challenges for companies in data processing. To address this, companies are turning to digital solutions for processing raw data and transforming it into visualized easily analyzable information. By applying machine learning techniques, both descriptive and predictive analyses are conducted to identify historical sales patterns, while predictive analysis leverages historical data to forecast future revenue trends. This study uses machine learning, specifically the K-Nearest Neighbors (KNN) and ARIMA models for data analysis. The integration of machine learning and Business Intelligence through real-time visualizations using tools like Power BI helps companies monitor performance metrics, identify opportunities, and respond to market changes effectively. This combination enables a more granular analysis of factors such as customer preferences, demand fluctuations, and supply chain efficiencies, which ultimately helps in tailoring strategies that maximize operational effectiveness and profitability. Additionally, machine learning models can be providing organizations with the agility needed to stay competitive in a dynamic marketplace to support the continuous improvement of business operations, resource allocation, and overall strategic planning.