In a movie, the story and the media are subordinate to each other; the story and images interact to form a narrative that is delivered to the audience, which has been central to the tradition in film criticism. This necessitates an integrated research on image representation and narrative. In this work, we study the quantitative difference between screenplay per se and the video using network analysis to highlight the translation process, since the relationships between characters form the central feature of the narrative. The two graphs showed a clear difference, so further investigation was implemented to clarify the cause of the difference using shot scale and facial expression data. We employ natural language processing and machine learning techniques for face recognition to do so. The results show the elaborate process by which a character is emphasized via image craft.