Evaluating Transitions for Streaming Big Data

Abstract

Visualizations for Streaming Big Data convey high volumes of information in real-time, making it challenging for people to grasp significant data changes. One solution could be having visualizations that change themselves according to the incoming data. However, these changes would need to be effectively conveyed. In this work, we propose a set of transitions between different pairs of visual idioms, aiming to aid users in keeping track of the information in real-time and notice relevant changes. We target transitions between Line charts, Heat maps, and Stream graphs. We conceived seven transitions that modify different properties of the visual elements for each pair of visual idioms, following a novel taxonomy for their conceptualization. To assess the performance of the transitions, we conducted an online user study with 100 participants. Results suggest that animations are indeed better to change between different visualization idioms than abrupt transitions. We also suggest transition techniques for each visualization pair, between those proposed, according to participants' preferences. Lastly, we identify which concepts of our taxonomy were more present in our suggested transitions.