This work proposes a traffic event model to assist the analysis of traffic events on big cities. This paper aims to provide not only an ontology for traffic events considering published news over the Internet, but also a prototype of a software architecture that uses the proposed model to perform queries on the events, using a huge quantity of published data on the Internet by regular users, especially on twitter. To do so, the news published in natural language is processed, and the relevant entities in the text are identified and structured in order to make a semantic analysis of them. The news reported is structured in the proposed model of events and thus the queries about their properties and relationships could be answered. As a consequence, the result of this work facilitates the analysis of the events occurred on the traffic process.