Local administrations face many challenges in their shift to smart city development. One of the main challenges is in the area of governance and identification of policy strategies that ensure successful integration of new technologies. In this article we explore a methodology for knowledge elicitation from textual information to identify smart city development trends. Our goal is to evaluate and predict the relevance of documents, especially press releases, to different areas of smart city development. Our methodology includes natural language processing techniques to possibilistically evaluate the relevance of a text to several smart city development areas. Relevant concepts are further associated to domain ontologies and expanded using data fusion techniques to evaluate the relevance of information in text to the areas of smart city development. We test our approach on a large number of documents from the smart city development community.