Several open government initiatives have been launched to make available online data enhancing accountability of public officials towards ordinary citizens. However, these initiatives raise several questions, namely: Which data should be disclosed? How to bring together dispersed (fragmented) data? How to improve its understandability by ordinary citizens? Literature shows that, in general, the data selection process does not take into account ordinary citizens' expectations and information needs. This paper presents the development process of a transparency ontology, which aims to provide an answer to the above questions, in what concerns public sector entities' use of resources. The process started by creating a list of relevant expressions/terms discussed in national and local newspapers, considering the role of journalists as 'information brokers' acting on behalf of ordinary citizens. This list was externally validated for relevance, comprehensiveness and improvements by interviewing journalists, and the resulting transparency ontology was formalised using OWL and Protégé.