Electronic participation data has a great potential for its application in studies of urban processes and systems. The popularity of e-participation services, which citizens use to report problems of urban improvement, housing and communal services, led to formation of massive datasets describing civic activity and subjective evaluation of urban environment quality. E-participation data has certain features that create distortions in the results of some research, e.g. when indirect evaluation of environmental or socioeconomic characteristics is concerned. One of the sources of such distortions are superusers. They are a small group of users whose activity is abnormally high. This abnormal activity has a significant impact on the distribution of messages. The fact that activity of e-participation service users is not equally distributed is well-known, however, a universal method of excluding the abnormal activity of superusers has never been proposed. This paper studies the distribution of activity of e-participation service users and proposes several methods of identifying and excluding peaks of increased activity of superusers in different territories and time intervals. The proposed methods were tested on data from the Russian portal “Our Saint Petersburg”. As a result, we have defined the optimal approaches to processing e-participation datasets for studies which are sensitive to the unequal distribution of subjective data.