The DYNAMAP project is expected to provide the traffic noise mapping of the city of Milan with an updating interval down to 5 minutes, thanks to a network of real-time noise monitoring stations. As the noise displayed by the maps should be exclusively attributed to the traffic noise, the recorded data need to be analyzed before the map updating process in order to exclude the presence of possible non-traffic events (e.g. sirens, horns, speech, doors, aircrafts,...), which are denoted as Anomalous Noise Events (ANE). To that effect, an Anomalous Noise Event Detection algorithm (ANED) has been designed and adjusted to deal with the identification of ANE within the Milan pilot area, where the mapping system is expected to be in operation in 2017. In this work, we present a comparative analysis of the ANE identification and classification considering the output of the ANED and the results obtained from the manual recognition of ANE by skilled personnel. The comparative takes into account the typology and duration of ANE (with their noise level distribution). In addition, their influence on the measured Leq,5min levels is evaluated for both methods by statistically analyzing the differences between the processed (without ANE) and non-processed (with ANE) Leq,5min.