Sales analysis using predictive analytics: a case study of Andara Heritage Spa, Kuala Terengganu

In the increasingly competitive wellness industry, small and medium enterprises like Andara Heritage Spa in Kuala Terengganu often struggle to leverage their historical data for strategic planning, relying instead on manual records and subjective intuition. This research addresses this gap by implementing the CRISP-DM framework to develop a predictive analytics system that transforms raw Point-of-Sale (POS) data into actionable business intelligence. Through a comparative analysis of machine learning algorithms in RapidMiner, the Random Forest model was identified as the most accurate predictor, achieving a Root Mean Squared Error (RMSE) of 54.612, which outperformed both k-NN and Decision Tree models in handling the nuances of spa transaction data. These predictive results were then integrated into an interactive Power BI dashboard, providing management with a visual tool to monitor performance and anticipate future revenue trends. Validation through domain expert evaluation confirmed that the system is both practically useful and strategically relevant, enabling the spa to transition from reactive management to proactive, data-driven decision-making for long-term growth.