[Purpose/significance] It is crucial for the emerging research field of data storytelling to define the conceptof data storytelling, to indetify the differences between data story and literary story as well as to implement data storyproducts automatically by software engineering methods. [Mothod/process] First, the concept and the features of datastorytelling are introduced respectively, based upon conducting acomprehensive literature review on the related studies. Second, the automatic generation process of data stories is proposed by data science principles. Third, a referenceframework for data story softeware enginneing is provided in line with Software Engineering methods. [Result/conclusion] The automatic generation process of data stories is of great significance for theoretical research on data storytelling, and its software engineering reference framework provides an integrated solution for developing data story productas well as nurturing its ecosystem.