The concept of smart cities has steadily increased in popularity over the last decade. This popularity has encouraged researchers to study the application of the concept to different aspects of a city. Consequently, the smart city literature is fraught with fragmented research contributions, partly because of a lack of consensus upon what makes a city smart. Recently, researchers have attempted to provide a coherent structure to the smart city literature in part to aid practitioners and policy makers in implementing the concept successfully. By reviewing the existing literature, studies have provided insights into research trends and progress. Nonetheless, they fall short on bringing together the fragmented research. Therefore, a novel conceptual model can help to organize the fragmented research and in translating the research progress into practice. To that end, we are proposing a conceptual model specific to smart city research in urban infrastructure. In this work, we have reviewed abstracts from the smart city literature and used Natural Language Processing (NLP) methods and clustering analysis to identify main research themes. Based on the analysis, we found the following 5 clusters: sustainability, smart energy, smart transportation, smart data transmission, and smart computing infrastructures. In our work, information from each cluster is incorporated into developing the conceptual model. We propose the “urban intelligence” conceptual model as a sequential model consisting of four stages: Acquire, Model, Analyze, Recommend. The main advantage of this model is that most of the relevant research contributions can be organized under each stage. This model also resembles the different phases in realizing a goal. Further, the model can be extended or linked to other aspects of a city. A notable contribution of this model is its ability to organize the existing research literature and to facilitate future research in smart city.