IEEE Transactions on Visualization and Computer Graphics, 2007
In interfaces that provide multiple visual information resolutions (VIR), low-VIR overviews typic... more In interfaces that provide multiple visual information resolutions (VIR), low-VIR overviews typically sacrifice visual details for display capacity, with the assumption that users can select regions of interest to examine at higher VIRs. Designers can create low VIRs based on multi-level structure inherent in the data, but have little guidance with single-level data. To better guide design tradeoff between display capacity and visual target perceivability, we looked at overview use in two multiple-VIR interfaces with high-VIR displays either embedded within, or separate from, the overviews. We studied two visual requirements for effective overview and found that participants would reliably use the low-VIR overviews only when the visual targets were simple and had small visual spans. Otherwise, at least 20% chose to use the high-VIR view exclusively. Surprisingly, neither of the multiple-VIR interfaces provided performance benefits when compared to using the high-VIR view alone. However, we did observe benefits in providing side-by-side comparisons for target matching. We conjecture that the high cognitive load of multiple-VIR interface interactions, whether real or perceived, is a more considerable barrier to their effective use than was previously considered.
Large data sets are difficult to analyze. Visualization has been proposed to assist exploratory d... more Large data sets are difficult to analyze. Visualization has been proposed to assist exploratory data analysis (EDA) as our visual systems can process signals in parallel to quickly detect patterns. Nonetheless, designing an effective visual analytic tool remains a challenge. This challenge is partly due to our incomplete understanding of how common visualization techniques are used by human operators during analyses, either in laboratory settings or in the workplace. This thesis aims to further understand how visualizations can be used to support EDA. More specifically, we studied techniques that display multiple levels of visual information resolutions (VIRs) for analyses using a range of methods. The first study is a summary synthesis conducted to obtain a snapshot of knowledge in multiple-VIR use and to identify research questions for the thesis: (1) low-VIR use and creation; (2) spatial arrangements of VIRs. The next two studies are laboratory studies to investigate the visual m...
Large-scale session log analysis typically includes statistical methods and detailed log examinat... more Large-scale session log analysis typically includes statistical methods and detailed log examinations. While both methods have merits, statistical methods can miss previously unknown subpopulations in the data and detailed analyses may have selection biases. We therefore built Session Viewer, a visualization tool to facilitate and bridge between statistical and detailed analyses. Taking a multiple-coordinated view approach, Session Viewer shows multiple session populations at the Aggregate, Multiple, and Detail data levels to support different analysis styles. To bridge between the statistical and the detailed analysis levels, Session Viewer provides fluid traversal between data levels and side-by-side comparison at all data levels. We describe an analysis of a large-scale web usage study to demonstrate the use of Session Viewer, where we quantified the importance of grouping sessions based on task type.
Two studies evaluated linear and hierarchy+elision small-screen display formats for clinical reas... more Two studies evaluated linear and hierarchy+elision small-screen display formats for clinical reasoning tasks. A controlled, quantitative study with 28 medically naive participants using a task abstracted from clinical use of laboratory results found that both display formats supported rapid and accurate decision making. Distribution of the search targets significantly affected speed, with decisions in linear format made 13% faster (4.7 sec) when all targets could be viewed on a single screen than when targets required scrolling between several screens and in hierarchical format 15 % faster (5.1 sec) when all the targets were confined within one category. Performance was equivalent regardless of the relative order of the target results and data in the laboratory report. In a qualitative study, 7 physicians used the displays to perform a realistic diagnosis. Physicians were comfortable with both display formats, but preference varied with clinical experience. The 5 less experienced cl...
Studied the visualization of large data sets to support exploratory data analysis Awarded the NSERC Postgraduate Scholarship (PGS-D2, 42K CDN)
Designed and implemented a medical laboratory result reporting system for handheld devices Gradua... more Designed and implemented a medical laboratory result reporting system for handheld devices Graduated with an A+ average
ABSTRACT MusicLand: Exploratory Browsing in Music Space
Most existing search tools based on query terms focus on direct search activities, where users ar... more Most existing search tools based on query terms focus on direct search activities, where users are assumed to have a clear and precise idea about their search targets. However, if users are uncertain about their targets, they will need to define or refine them with a succession of multiple related queries. Current search tools do not adequately support this process. MusicLand is designed for exploratory visual browsing of digitized music collections annotated with metadata. It supports the gradual refinement of queries by showing visually structured search results connected with a query history trail, and providing suggestions for new query terms. Semantic zooming maximizes information density, and animated transitions between display states help the users maintain their mental map of the relationship between new and old queries. 1
In order to display web pages designed for desktop-sized monitors, some small-screen web browsers... more In order to display web pages designed for desktop-sized monitors, some small-screen web browsers provide singlecolumn or thumbnail views. Both have limitations. Singlecolumn views affect page layouts and require users to scroll significantly more. Thumbnail views tend to reduce contained text beyond readability, so differentiating visually similar areas requires users to zoom. In this paper, we present Summary Thumbnails—thumbnail views enhanced with readable text fragments. Summary Thumbnails help users identify viewed material and distinguish between visually similar areas. In our user study, participants located content in web pages about 41 % faster and with 71% lower error rates when using the Summary Thumbnail interface than when using the Single-Column interface, and zoomed 59 % less than when using the Thumbnail interface. Nine of the eleven participants preferred Summary Thumbnails over both the Thumbnail and Single-Column interfaces. ACM Classifiction: H5.2 [Information i...
Despite the long history and consistent use of quantitative empirical methods to evaluate informa... more Despite the long history and consistent use of quantitative empirical methods to evaluate information visualization techniques and systems, our understanding of interface use remains incomplete. While there are inherent limitations to the method, such as the choice of task and data, we believe the utility of study results can be enhanced if they were amenable to meta-analysis. Based on our experience in extracting design guidelines from existing quantitative studies, we recommend improvements to both study design and reporting to promote meta-analysis: (1) Use comparable interfaces in terms of visual elements, information content and amount displayed, levels of data organization displayed, and interaction complexity; (2) Capture usage patterns in addition to overall performance measurements to better identify design tradeoffs; (3) Isolate and study interface factors instead of overall interface performance; and (4) Report more study details, either within the publications, or as sup...
Research in face-to-face communication has identified a variety of largely unconscious patterns o... more Research in face-to-face communication has identified a variety of largely unconscious patterns of behavior that are used by collaborators to coordinate their conversation, build common ground, and repair errors of understanding. These provide a communication metachannel that is not explicitly supported in current collaboration environments. Without this metachannel, conversants must overburden linguistic channels with explicit communication strategies and codified behaviours e.g. "netiquette". One approach to solving this problem is to attempt to reproduce face-to-face collaboration with VR, video, etc. We have taken an alternative approach, using visualization techniques to create a graphical network representation of patterns of reference in collaborative discourse. A preliminary user study suggested that the CZTalk proof-of-concept prototype provided improvements in both interpretation and authorship.
Query-level search intent categorizations alone cannot capture the complexity that spans search t... more Query-level search intent categorizations alone cannot capture the complexity that spans search tasks. We come to this conclusion after an extensive analysis of real users labeling their own web search tasks. Analyzing this data, we derive a search-intent classification that is task-based (rather than single-query-based) and far more descriptive than other classifications. To derive the taxonomy, we conducted a diary study with 36 participants for two weeks. Each participant self-segmented their web activity stream into tasks and labeled each task with their intent. We used data from 34 participants spanning 1463 web search tasks and ran an affinity diagramming exercise, grouping tasks by intent. The result was a twelve category categorization of task intent organized along two main axes: (a) specificity: degree of specificity of the user’s desired end goals, and (b) decision- related: whether the user was in the process of making a decision based on information found during the task.
We examined the effects of geometric transformations and their interactions with background grids... more We examined the effects of geometric transformations and their interactions with background grids on visual memory to provide interface design guidelines. We studied scaling, rotation, rectangular fisheye, and polar fisheye transformations. Based on response time and accuracy results, we defined a no-cost zone for each transformation type within which performance is unaffected. Results indicated that scaling had no effect down to at least 20% reduction. Rotation had a no-cost zone of up to 45 degrees, after which the response time increased to 5.4 s from the 3.4 s baseline without significant drop in accuracy. Interestingly, polar fisheye transformation had a lesser effect on accuracy than the rectangular counterpart. The presence of grids extended these zones and significantly improved accuracy in all but the polar fisheye transformation trials. Our results therefore provided guidance on the types and levels of nonlinear transformations that could be used without affecting performa...
Google Fusion Tables (GFT) brings big data collaboration and visualization to data-experts who po... more Google Fusion Tables (GFT) brings big data collaboration and visualization to data-experts who possess neither large data-processing resources nor expertise. In this paper we highlight our support for map visualizations over large complex geospatial datasets. Interactive maps created using GFT have already been used by journalists in numerous high-profile stories.
Abstract—Interaction cost is an important but poorly understood factor in visualization design. W... more Abstract—Interaction cost is an important but poorly understood factor in visualization design. We propose a framework of interaction costs inspired by Norman’s Seven Stages of Action to facilitate study. From 484 papers, we collected 61 interaction-related usability problems reported in 32 user studies and placed them into our framework of seven costs: (1) Decision costs to form goals; (2) System-power costs to form system operations; (3) Multiple input mode costs to form physical sequences; (4) Physical-motion costs to execute sequences; (5) Visual-cluttering costs to perceive state; (6) View-change costs to interpret perception; (7) State-change costs to evaluate interpretation. We also suggested ways to narrow the gulfs of execution (2–4) and evaluation (5–7) based on collected reports. Our framework suggests a need to consider decision costs (1) as the gulf of goal formation.
Large-scale session log analysis typically includes statistical methods and detailed log examinat... more Large-scale session log analysis typically includes statistical methods and detailed log examinations. While both methods have merits, statistical methods can miss previously unknown subpopulations in the data and detailed analyses may have selection biases. We therefore built Session Viewer, a visualization tool to facilitate and bridge between statistical and detailed analyses. Taking a multiple-coordinated view approach, Session Viewer shows multiple session populations at the Aggregate, Multiple, and Detail data levels to support different analysis styles. To bridge between the statistical and the detailed analysis levels, Session Viewer provides fluid traversal between data levels and side-by-side comparison at all data levels. We describe an analysis of a large-scale web usage study to demonstrate the use of Session Viewer, where we quantified the importance of grouping sessions based on task type.
Google Fusion Tables (GFT) brings big data collaboration and visualization to data-experts who po... more Google Fusion Tables (GFT) brings big data collaboration and visualization to data-experts who possess neither large data-processing resources nor expertise. In this paper we highlight our support for map visualizations over large complex geospatial datasets. Interactive maps created using GFT have already been used by journalists in numerous high-profile stories.
A Study-Based Guide to Multiple Visual Information Resolution Interface Designs
ed and visually abbreviated to increase the display capability of the low VIRs. Section 5.3 discu... more ed and visually abbreviated to increase the display capability of the low VIRs. Section 5.3 discusses cases where the designers have gone too far in their abstraction, and the study participants could no longer use the visual information on the low VIRs. Instead of abstraction, the designer could choose to selectively display or emphasize a subset of the data in the low VIRs, for example, based on the generalized fisheye Degree of Interest function [Furnas 1986]. However, study result suggest that a priori automatic filtering is a double-edged sword, as discussed in Section 5.4. Given all these consideration, we round up the discussion by re-examining the roles of low VIRs in Section 5.5 to help ground the low VIR design. DECISION 3 (Section 6): Simultaneous or temporal display of the VIRs. Once the low VIRs are created, the designer would then need to display them, either simultaneously as in the embedded or the separate interfaces, or one VIR at a time as in the temporal interface...
Proceedings of the Beyond Time and Errors on Novel Evaluation Methods for Visualization - BELIV '16
Visualization has shown its ability to produce powerful tools for analyzing, understanding, and c... more Visualization has shown its ability to produce powerful tools for analyzing, understanding, and communicating data and making it accessible for several different tasks and purposes. Impact of visualization to everyday work and personal lives is demonstrated by many successes stories---such as the increasing prevalence of Tableau, the interactive visualizations produced by the New York Times, or toolkits like VTK/Paraview to name just a few. A large community of casual and professional users are increasingly consuming and producing both interactive and static visualizations.
Most existing search tools based on query terms focus on direct search activities, where users ar... more Most existing search tools based on query terms focus on direct search activities, where users are assumed to have a clear and precise idea about their search targets. However, if users are uncertain about their targets, they will need to define or refine them with a succession of multiple related queries. Current search tools do not adequately support this process. MusicLand is designed for exploratory visual browsing of digitized music collections annotated with metadata. It supports the gradual refinement of queries by showing visually structured search results connected with a query history trail, and providing suggestions for new query terms. Semantic zooming maximizes information density, and animated transitions between display states help the users maintain their mental map of the relationship between new and old queries.
Figure 1: Views depicting reported cases of a disease across space and time. Even for this small ... more Figure 1: Views depicting reported cases of a disease across space and time. Even for this small set of views, many sequences are possible.
IEEE Transactions on Visualization and Computer Graphics
Visualization researchers and practitioners engaged in generating or evaluating designs are faced... more Visualization researchers and practitioners engaged in generating or evaluating designs are faced with the difficult problem of transforming the questions asked and actions taken by target users from domain-specific language and context into more abstract forms. Existing abstract task classifications aim to provide support for this endeavour by providing a carefully delineated suite of actions. Our experience is that this bottom-up approach is part of the challenge: low-level actions are difficult to interpret without a higher-level context of analysis goals and the analysis process. To bridge this gap, we propose a framework based on analysis reports derived from open-coding 20 design study papers published at IEEE InfoVis 2009-2015, to build on the previous work of abstractions that collectively encompass a broad variety of domains. The framework is organized in two axes illustrated by nine analysis goals. It helps situate the analysis goals by placing each goal under axes of specificity (Explore, Describe, Explain, Confirm) and number of data populations (Single, Multiple). The single-population types are Discover Observation, Describe Observation, Identify Main Cause, and Collect Evidence. The multiple-population types are Compare Entities, Explain Differences, and Evaluate Hypothesis. Each analysis goal is scoped by an input and an output and is characterized by analysis steps reported in the design study papers. We provide examples of how we and others have used the framework in a top-down approach to abstracting domain problems: visualization designers or researchers first identify the analysis goals of each unit of analysis in an analysis stream, and then encode the individual steps using existing task classifications with the context of the goal, the level of specificity, and the number of populations involved in the analysis.
Uploads
Papers by Heidi Lam