The model of an intelligent data analyzer of city logistics, transportation, social and economic behavior of people and the state objects of urban systems has been developed in this paper. Personalized recommendation control elements of the urban system have been described. The concept of a directive element of the urban system was given and defined. The problem of prevention of causes of accidents has been detailed. Searching for causes of accidents was solved through an abductive inference process. The Codified Template Library of Potentially Dangerous Patterns of Behavior was created with an inductive inference process. Some examples and recommendations for improvement of the GATLOSEMI model were given.