Drillboards: Adaptive Visualization Dashboards for Dynamic Personalization of Visualization Experiences

TL;DR

This paper introduces drillboards, an adaptive visualization framework that represents dashboards as a hierarchical hierarchy of charts enabling drill-down and roll-up interactions. The DrillVis authoring tool supports constructing the aggregation hierarchy and predefined views, while a reader mode lets end users navigate tailored views matching their expertise and tasks. The approach combines a formal vocabulary of chart representations with merge operations (Label, Summarization, Archetype, Projection, Juxtaposition, Overlay) to build scalable, personalized views without expanding screen space. A four-phase user study with three domain experts and ten casual end-users demonstrates that drillboards aid expert storytelling and novice understanding, highlighting potential for improved communication and learning in data-rich contexts.

Abstract

We present drillboards, a technique for adaptive visualization dashboards consisting of a hierarchy of coordinated charts that the user can drill down to reach a desired level of detail depending on their expertise, interest, and desired effort. This functionality allows different users to personalize the same dashboard to their specific needs and expertise. The technique is based on a formal vocabulary of chart representations and rules for merging multiple charts of different types and data into single composite representations. The drillboard hierarchy is created by iteratively applying these rules starting from a baseline dashboard, with each consecutive operation yielding a new dashboard with fewer charts and progressively more abstract and simplified views. We also present an authoring tool for building drillboards and show how experts users can use to build up and deliver personalized experiences to a wide audience. Our evaluation asked three domain experts to author drillboards for their own datasets, which we then showed to casual end-users with favorable outcomes.

Figures (13)

• Figure 1: Drillboards are adaptive visualization dashboards consisting of a hierarchy of coordinated charts. A drillboard has two modes: authoring (A) and reading (B). The hierarchy is designed in authoring mode, and then explored in reading mode. The author-designed hierarchical structure is shown in the hierarchy browser. In the reading mode, the user can drill down a chart to expand it into its constituent child charts. The user can also roll up the hierarchy to collapse a few charts. Here is how the author mode (A) works: ① The author needs to select charts of their interest. ④ After selecting the operation to merge the charts and adding descriptions and title, ⑤ the chart is merged. The treeview also is updated based on the changes made (②, ③). Here is how the reader mode (B) works: ① The reader first sees a chart of their interest, and left-clicks the chart. Then, ⑥ the original chart disappears, and 3 children charts appear instead, in a highlighted manner. ⑤ By right-clicking any children charts, the charts roll-up to transform into the parent chart. Here too, the treeview (①, ③) is updated based on the changes made in the chart view.
• Figure 2: Author and reader modes in DrillVis. In the author mode, authors can create aggregate hierarchies using unit charts. In reader mode, users can drill down into visualizations to explore the details hidden beneath the selected chart.
• Figure 3: Abstract aggregation operations. Schematic illustrations of the abstract aggregate operations that can be applied to charts in a drillboard to build the aggregation hierarchy: \ref{['fig:agg-label']} labeling the aggregate with text to represent the children; \ref{['fig:agg-summarize']} summarizing the children with a data abstraction (e.g., mean); \ref{['fig:agg-archetype']} selecting one child as archetype to represent them all; \ref{['fig:agg-project']} projecting child data onto axes to form a scatterplot or parallel coordinate plot; \ref{['fig:agg-juxtapose']} juxtaposing multiple charts in the aggregate; and \ref{['fig:agg-overlay']} overlaying multiple data series in the same chart.
• Figure 4: Hierarchical aggregation. Example of how an aggregation hierarchy is formed through a sequence of aggregation operations until a single pile remains as the root.
• Figure 5: Visualization cards. DrillVis uses cards as containers for visualizations. Single cards (left) represent a single chart atom ; piled card visualization (right) represent piles .