Urban thermal modeling is often constrained by the limited availability of high-resolution spatiotemporal data. In London’s Queen Elizabeth Olympic Park, we deployed a network of 15 bespoke IoT temperature sensors across various land cover types. We compared the measured air temperature (Tair) with UMEP modeled mean radiant temperature (Tmrt). A rank analysis indicated a significant positive correlation for daytime which confirms the sensors’ ability to resolve microclimate variations, but a weaker correlation for nighttime, indicating limitations of current thermal modelling methods. Results demonstrate the value of the cost-effective IoT sensors in detecting, monitoring and remediating thermal hotpots.