Page 177 - Journal of Library Science in China, Vol.45, 2019

176
176 Journal of Library Science in China, Vol.11, 2019

cognition.
Data storytelling involves three basic elements: data, visuals, and narrative. Combining data and
visuals can enlighten the audience to gain data insights; applying narrative into data can explain
the data phenomena; coupling visuals with narrative can entertain the audience. The main activity
flow of data story-telling is divided into six steps: 1) The author must understand the data and
make clear the meaning it represents. 2) The author who makes data stories needs to have a clear
purpose. 3) Identifying the audiences as novices, generalists, managers, experts or executives
helps author create different data stories for different audiences. 4) The next step is that the author
identifies key data and uses the most effective data to describe the data story. 5) Then the author
chooses the story model, chooses the appropriate chart, and describes the information presented
by the story according to people’s visual characteristics. 6) Providing background narration and
guiding the audience according to the plot effectively synthesizes and organizes data stories. In
addition, the following three special activities may be involved in the actual data storytelling
project: 1) Data story-telling experiment and pre-investigation. 2) Continuous improvement of data
storytelling. 3) Separation between the author and the narrator of the data story.
It is worth mentioning that the transformation from data to story and the presentation of data
story are two different stages of data storytelling. The author of data story needs to convert the data
into a story model beforehand, and then the narrator takes a special way to present the story. There
are two presentation forms of data stories: 1) The narrator tells the story and the audience listens to
the story. 2) The narrator shows the data story to the audience, and the audience can see the story.
Data storytelling is one of the main research contents of data science, and it is also an important
feature that distinguishes data science from other disciplines. The application of data storytelling
in data science is mainly reflected in four aspects. First, data storytelling solves the “last mile”
problem of data science, which plays a crucial role in the success of data science projects; second,
data storytelling is an important means of obtaining insights from big data; third, data storytelling
can transform data insights into data actions; fourth, data storytelling is an important activity of
data product development, such as data journalism.
Finally, this paper summarizes five main characteristics of the current research: 1) Foreign
research is more than domestic research. 2) Research articles published informally (such as
blogs) are more than the officially published academic papers. 3) Theoretical research lags
behind the practice application. 4) There is more research on concept level than the realization
of concrete technology. 5) The visualization tools for data story are more than specialized tools
of data storytelling. In view of these characteristics, some recommended topics in the following
research of data storytelling are proposed: 1) Improving the theoretical system of data storytelling.
2) Studying the evaluation method of data storytelling, optimizing specific data story projects.
3) Strengthening interdisciplinary research and further broadening the research perspective and
theoretical basis of data storytelling. 4) Exploring new algorithms and models for data storytelling.
5) Developing special tools for data storytelling.