Humans constantly create narratives to provide explanations for how and why something happens. Sherlock Holmes is known for his observation and logical reasoning skills, and is called upon for finding the identity and motivations of the culprits. In other words, we constantly attempt to make sense of different inputs and to come up with a coherent story. In my research, I investigate how to computationally build structured narratives with knowledge graphs as inputs. The objective is to find and implement a meaningful knowledge representation, and to find relevant inputs for the output graph. I am particularly interested in dynamic representations and in reasoning over the sequence of events’ patterns. By adding the reasoning step, we will be able to make new hypotheses and, therefore, to discover new knowledge.