This paper presents an algorithm to characterize power quality (PQ) events. Any power quality event captured at the monitoring point will be analyzed automatically in three stages. The first stage differentiates between sag, swell or transient events by using wavelet analysis as well as time domain analysis for counter checking purposes. Any sag events will automatically be passed to second stage to identify if the sag is caused by induction motor starting, transformer energizing or line faults. Important parameters such as phase angle shift, non/symmetrical sag, non/rectangular sag and sag duration are extracted and used for event characterization in the second stage. If a line fault is identified in the second stage, pseudo-measurement and transformer modeling will be executed to identify the fault location together with the 3-phase bus voltages at the faulted point. Here, the third stage is triggered and the line fault is further characterized into single line-to- ground fault, line-to-line fault or double line-to-ground fault by using the zero sequence voltage magnitude as well as zero sequence voltage angle. This automated event characterization algorithm provides a fast identification of areas that are prone to certain types of faults, which leads to much easier identification of causes of faults in that area - such as trees, animals, nature and etc. Preventive measures as well as the correct mitigation option can then be designed to reduce such incidences and thus improves power quality.