Exploring Visualization of Data Transforms

Abstract

In the context of data exploration, users often interact with relational database systems in an interactive query session to form useful insights. Each query a user executes can potentially transform a resultset in complex ways. We explore some of the challenges in understanding these transformations, and how these challenges can be solved through more informative visual representations of data transforms. We present the concept of "tweening" of resultsets as a method of incrementally visualizing data transformations, and explore approaches towards generating these resultset tweens. Through a series of user studies, we evaluate tweening as an effective method of understanding the changes that result from data transformations.