With rising summer temperatures causing significant health hazards, understanding fine-scale temperatures patterns is becoming increasingly important. Public urban IoT sensor networks have the potential to provide such information; however, the high cost of these networks represents a considerable constraint, particularly for smaller municipalities. Therefore, this study aimed to quantitatively evaluate to quantitatively evaluate their effectiveness by using the high-density public urban IoT sensor network (S-DoT) installed by the Seoul Metropolitan Government as a case study. The difference in daily maximum temperature (ΔTX) between conventional reference stations and S-DoT was quantified, and its relationships with variables representing urban landscape and structure were examined. Key findings include: (1) ΔTX ranged from -0.93°C to +2.36°C. (2) Positive ΔTX (S-DoT > reference) trend was more prominent in compact low- to mid-rise urban areas near green spaces. (3) Negative ΔTX (S-DoT < reference) trend was stronger in areas with forest land cover on steep-slope, likely due to the cooling effect of tree shade. (4) Large ΔTX values were also derived by the prediction uncertainty of the reference station network. This study demonstrates that S-DoT sensors can identify critical thermal risk areas in compact low- to mid-rise urban zones near green spaces that cannot be captured by conventional reference stations. In addition, it provides insights into the strategic placement of IoT sensors to bridge undetected temperature gaps between vertical levels, thereby offering valuable information for decision-makers to strengthen urban heat risk assessment and adaptation strategies.