Animated narrative visualization for video clickstream data

  • ,
  • Zhutian Chen ,
  • Quan Li ,
  • Xiaojuan Ma ,
  • Qiong Luo ,
  • Huamin Qu

SA '16: SIGGRAPH ASIA 2016 Symposium on Visualization |

Published by ACM

Publication | Publication | Publication | Related File

Video clickstream data are important for understanding user behaviors and improving online video services. Various visual analytics techniques have been proposed to explore patterns in these data. However, those techniques are mainly developed for analysis and do not sufficiently support presentations. It is still difficult for data analysts to convey their findings to an audience without prior knowledge. In this paper, we propose to use animated narrative visualization to present video clickstream data. Compared with traditional methods which directly turn click events into animations, our animated narrative visualization focuses on conveying the patterns in the data to a general audience and adopts two novel designs, non-linear time mapping and foreshadowing, to make the presentation more engaging and interesting. Our non-linear time mapping method keeps the interesting parts as the focus of the animation while compressing the uninteresting parts as the context. The foreshadowing techniques can engage the audience and alert them to the events in the animation. Our user study indicates the effectiveness of our system and provides guidelines for the design of similar systems.