Annotation Graphs: A Graph-Based Visualization for Meta-Analysis of Data Based on User-Authored Annotations

Abstract

User-authored annotations of data can support analysts in the activity of hypothesis generation and sensemaking, where it is not only critical to document key observations, but also to communicate insights between analysts. We present annotation graphs, a dynamic graph visualization that enables meta-analysis of data based on user-authored annotations. The annotation graph topology encodes annotation semantics, which describe the content of and relations between data selections, comments, and tags. We present a mixed-initiative approach to graph layout that integrates an analyst's manual manipulations with an automatic method based on similarity inferred from the annotation semantics. Various visual graph layout styles reveal different perspectives on the annotation semantics. Annotation graphs are implemented within C8, a system that supports authoring annotations during exploratory analysis of a dataset. We apply principles of Exploratory Sequential Data Analysis (ESDA) in designing C8, and further link these to an existing task typology in the visualization literature. We develop and evaluate the system through an iterative user-centered design process with three experts, situated in the domain of analyzing HCI experiment data. The results suggest that annotation graphs are effective as a method of visually extending user-authored annotations to data meta-analysis for discovery and organization of ideas.

Supporting Handoff in Asynchronous Collaborative Sensemaking Using Knowledge-Transfer Graphs

“…Graph-based visualizations have been used extensively to externalize investigations in visual analysis systems, including investigating document collections (e. g., [23,28,45]), intelligence analysis (e. g., [3,7,29,35], and visual analytics (e.g., [6,55]). KTGraph builds on the general visual design of our previous work, Annotation Graphs [55], that offered specific exploration capabilities for graphs based on added data annotations. KTGraph, while visually similar, has a very different goal.…”

“…While in this paper we focus on document analysis tasks, KTGraph is generalizable by replacing the Dataset Panel with other visualizations for specific applications, such as the time-series data tables in Annotation Graphs [55]. Instead of representing text and text excerpts, references could then link to data chunks selected by an analyst.…”

“…8), and thus subsequent handoffs would become more and more difficult due to visual clutter. Existing methods, such as searching and filtering graph elements or methods for collapsing nodes and links [6,55] could address these scalability issues.…”

“…From our observations in the studies, participants did not tend to manually organize the graph "[...] unless I was forced to, because it takes too much effort." Thus, automatic or semi-automatic graph layout methods based on certain similarity measures among nodes, links, comments, and references (as in e. g., AnnotationGraphs [55]) are worth further consideration.…”

“…Yet, the features that support this task could also be applied to a variety of other exploratory data analysis scenarios where investigators generate hypotheses. For example, our tool can support workflows of collaborative visual analysis tools that involve selection and annotation of data points, such as for time-series data [55] or charts [13,52]. With an extension of LCW support like CLIP [29], KTGraph may be used in both synchronous and asynchronous collaborations.…”