-
-
First Page
Luck, Location, or a Leak? Dashboard Royale | TAMU Datathon 2026 Project Overview The Gilded Ace Casino faces a critical business decision: whether to replace its ten lowest-earning slot machines. However, low earnings do not necessarily indicate poor machine performance. Differences in location, availability, player demand, and payout luck can significantly influence revenue. Our project uses Power BI, Python, and machine learning to investigate these factors and develop a data-driven investment strategy. Analysis & Key Findings We analyzed January–June 2025 casino operations, evaluating machine profitability, expected house win, wagering activity, downtime, and tracked-player behavior. Our analysis revealed that 9 of the 10 lowest-earning machines were located in Back Forty, a zone generating approximately 45% less expected house win per available hour than Main Entrance. This suggests that floor location and player exposure may contribute to poor performance, making immediate machine replacement potentially premature. We also compared actual and theoretical house win to distinguish payout luck from underlying performance. Across the casino, actual house win totaled approximately $23.64 million, compared with $23.55 million expected, producing a positive luck residual of approximately $92,790. Additional dashboard views examine loyalty-program costs, tracked-player outcomes, and estimated revenue opportunities associated with machine downtime. Predictive Modeling We developed a Random Forest model to forecast daily casino coin-in and compared it against a simple weekday-average baseline. Using a chronological training and testing split, with May–June 2025 reserved for evaluation, our model achieved a mean absolute error of $74,952, compared with $77,710 for the baseline—an improvement of approximately 3.55%. These forecasts provide an additional tool for anticipating demand and informing operational planning. Investment Recommendation Rather than immediately replacing underperforming machines, we recommend a targeted 30-day intervention: $7,200: Relocate four machines to evaluate location effects. $4,800: Inspect four machines for payout or mechanical issues. $11,520: Add technician coverage to address downtime. This allocates $23,520 of the $100,000 budget, preserving $76,480 until additional evidence justifies further spending. Conclusion Our findings support a measured, evidence-driven approach to casino management. By distinguishing machine performance from location, availability, and payout luck, The Gilded Ace can test lower-cost interventions before committing to expensive replacements. Success will be evaluated through changes in expected earnings per available hour, machine availability, and operational performance. I’m also disclosing that AI was used in this process.
