Often manually performed commissioning processes on building's sensors fail to systematically validate that all building's sensors operate correctly. This is so because manual processes are tedious and only inspect a limited number of sensors. As a result, sensors are often uncalibrated, biased or somehow faulty, impacting building's behaviour, comfort level and energy usage. We present a practical approach to automatically validate data from all building's sensors. We designed and implemented four different tests to detect out-of-range values, spikes, latency issues and non-monotonous values. Our tests are based on expert knowledge and do not need historical data. We ran the validation tests on a newly constructed building at the campus of the University of Southern Denmark. As a result we identified two types of faulty behaviours in the building's sensors: CO2 sensors reporting biased values and temperature sensors' readings exhibiting high latency. We show how automatic data validation for building sensors enhances the processes of detecting issues which could severely impact building's operations, and were otherwise going unnoticed. Thus, we emphasize the importance of performing data validation as a necessity for a correct building operation.