Over the last years there has been a relevant increase of large scale data. “Volume”, Velocity”, “Variety”, “Veracity” and, more importantly today, “Value” are the dimensions that characterize Big Data. Big data are increasingly important in the real world. In this paper, we deal with bibliometric datasets which represent relevant information and value for applications, usually difficult to identify. In economic analyses, bibliographic networks are useful to represent the data and provide relevant insights on research findings. In this work we will propose to apply a framework based on symbolic data and, in particular, one based on data-based symbolic observation interval which represent relevant patterns and information from complex data. These results are important because they consider a relevant case of complex information which is transformed into a representation useful to be analysed as network data. From these network representations of the information (considering a co-occurrence network from the relevant concepts of the studied literature of “regression discontinuity”) we are able to identify the most relevant patterns in data, as “communities” of concepts maximally connected to each other. From the communities we are able to represent the semantic cores of the literature.