This study aimed to estimate the risk of airborne SARS-CoV-2 transmission across 44 indoor clinical spaces in a rehabilitation hospital using environmental modeling and multivariate statistics. High-resolution CO₂ measurements, occupancy, and ventilation data were collected, and the Gammaitoni–Nucci model was used to compute individual transmission risk estimates. Six indoor air quality (IAQ) variables were measured: room volume, occupancy, CO₂ concentration, time with CO₂ > 1000 ppm, ventilation, and air replacement time. Among 36 valid observations, the mean CO₂ concentration was 1012.4 ppm (SD = 313.7), the average ventilation rate was 4.4 ACH, and the estimated transmission risk was 28.3% (SD = 18.9). Principal Component Analysis (PCA) revealed that two components explained 68.8% of total variance: PC1 captured ventilation efficiency (high ACH), while PC2 reflected occupancy density and elevated CO₂ levels. Hierarchical clustering identified three distinct IAQ profiles: Cluster 1 (high occupancy, high CO₂, moderate ventilation), Cluster 2 (low occupancy, poor ventilation, highest risk), and Cluster 3 (high ACH, low CO₂, low exposure). While traditional service-based comparisons showed no differences, cluster-based analysis revealed significant variations in CO₂ (p < 0.01), ACH (p < 0.01), occupancy (p < 0.01), and transmission risk (p = 0.01). Multiple linear regression identified ACH as the only significant predictor of airborne risk (β = −1.25, p = 0.01), resulting in a 1.25 percentage point reduction in risk per ACH unit. These findings support the use of IAQ-based clustering as a practical framework for identifying the risk of airborne transmission. Recommendations include maintaining ACH ≥3 and prioritizing CO₂ monitoring in high-turnover rooms.