Abstract: 【Purpose/significance】As a link between events and story, the plot is one of the pillars of data storytelling. It is important to discuss the theoretical research, automatic generation and engineering development of data story plot.【Method/process】Firstly, the con⁃ cept of data story plot and its difference from literary plot are defined, and the Freytag′s Pyramid structure used in this study is clarified. Secondly, the elements of plot design are discussed in terms of the events in the plot and the plot structure. and based on this, an auto⁃ matic plot generation and presentation method is proposed, mainly involving event recommendation, event sampling, plot visualisation and plot mapping. Then, the paper takes Freytag′s Pyramid structure as an example, draws on SP-LIME, information entropy decision tree and SMOTE algorithms to achieve data story plot generation. Finally, the paper shows the plot generation effect through a tree dia⁃ gram and a Hype Cycle graph.【Result/conclusion】For the first time in the field of data storytelling, types of data story plot and design el⁃ ements are proposed, and an automatic generation method for Freytag′s Pyramid structure is proposed.【Innovation/limitation】This study makes up for the lack of research at this stage on the constituent elements of data storytelling, proposes an experimentally feasible plot gen⁃ eration method and plot structure mapping ideas, and provides certain insights for the subsequent in-depth study of data story plot.