Graphical Histories for Visualization: Supporting Analysis, Communication, and Evaluation

Abstract

Abstract-Interactive history tools, ranging from basic undo and redo to branching timelines of user actions, facilitate iterative forms of interaction. In this paper, we investigate the design of history mechanisms for information visualization. We present a design space analysis of both architectural and interface issues, identifying design decisions and associated trade-offs. Based on this analysis, we contribute a design study of graphical history tools for Tableau, a database visualization system. These tools record and visualize interaction histories, support data analysis and communication of findings, and contribute novel mechanisms for presenting, managing, and exporting histories. Furthermore, we have analyzed aggregated collections of history sessions to evaluate Tableau usage. We describe additional tools for analyzing users' history logs and how they have been applied to study usage patterns in Tableau.Index Terms-Visualization, history, undo, analysis, presentation, evaluation. INTRODUCTIONWhen investigating data with visualizations, users regularly traverse the space of views in an iterative fashion. Exploratory analysis may result in a number of hypotheses, leading to multiple rounds of question-answering. Analysts can generate unexpected questions that may be investigated immediately or revisited later. After conducting analysis, users may need to review, summarize, and communicate their findings, often in the form of reports or presentations. By surfacing users' interaction history, we can facilitate analysis and communication. History mechanisms such as undo or "timetravel" enable revisitation in a variety of applications (e.g., [1-4, 6, 8, 12-14, 16-20, 22, 24-26]). As noted by Shneiderman [27], such history tools can play an important part in the visualization process, supporting iterative analysis by enabling users to review, retrieve, and revisit visualization states. Moreover, history tools can help users create reports or presentations, facilitating communication.Interaction histories can also benefit research and development. History log analysis of both individual and aggregate usage can identify common usage patterns and thereby assist usability evaluation. Researchers can also study interaction patterns to better understand and model analysts' sense-making process [13].However, the best history mechanisms for achieving these benefits are not always clear. Designers of visualization tools must consider a large design space of potential features and system architectures when designing history tools. These design decisions entail trade-offs in the types of history representations and operations that can be provided.For example, while it is easy to log low-level input events such as key presses and mouse clicks [25], users can more readily take advantage of semantically meaningful models. Many operations might be performed on an interaction history, including editing, aggregation, bookmarking, annotation, and search. Architecture and interface design need to account for such operations. ...

Video Navigation with a Personal Viewing History

“…education, entertainment, etc. History of user actions has been widely investigated for multiple purposes in different domains including web [2,19,15], workflow [12,21], tutorial generation [1] and information visualizations [14,13]. These are based on mostly static media and so do not directly apply to video: unlike static media, two timelines exist when capturing user history: video timeline and user (or history) timeline.…”

Hoptrees: Branching History Navigation for Hierarchies

“…For supporting visual analysis tasks, Heer et al describe a design space of graphical history tools and provide a review of recent research in this area [13]. History tracking widgets with branching structures have also been investigated for visual analytics tasks [14,15].…”

Interactive visual analysis promotes exploration of long‐term ecological data

“…2) including aggregated static data (for bins and line series) as opposed to providing dynamic access to the most current data (Heer et al 2008b). This implementation decision is based on the understanding of characteristics of large ecological data sets.…”