DATA-DRIVEN STORYTELLING FOR CASUAL USERS

Abstract

Today’s overwhelming volume of data has made effective analysis virtually inaccessible for the general public. The emerging practice of data-driven storytelling is addressing this by framing data using familiar mechanisms such as slideshows, videos, and comics to make even highly complex phenomena understandable. However, current data stories still do not utilize the full potential of the storytelling domain. One reason for this is that current data-driven storytelling practice does not leverage the full repertoire of media that can be used for storytelling, such as speech, e-learning, and video games.

In this dissertation, we propose a taxonomy focused specifically on media types for the purpose of widening the purview of data-driven storytelling by putting more tools in the hands of designers. We expand the idea of data-driven storytelling into the group of casual users, who are the consumers of information and non-professionals with limited time, skills, and motivation , to bridge the data gap between the advanced data analytics tools and everyday internet users. To prove the effectiveness and the wide acceptance of our taxonomy and data-driven storytelling among the casual users, we have collected examples for data-driven storytelling by finding, reviewing, and classifying ninety-one examples.

Using our taxonomy as a generative tool, we also explored two novel storytelling mechanisms, including live-streaming analytics videos—DataTV—and sequential art (comics) that dynamically incorporates visual representations—Data Comics. Meanwhile, we widened the genres we explored to fill the gaps in the literature. We also evaluated Data Comics and DataTV with user studies and expert reviews. The results show that Data Comics facilitates data-driven storytelling in terms of inviting reading, aiding memory, and viewing as a story. The results also show that an integrated system as DataTV encourages authors to create and present data stories.