Chair: Jean-Daniel Fekete, INRIA, France
Interpretation and Trust: Designing Model-Driven Visualizations for Text Analysis
Contribution & Benefit: Proposed criteria (interpretation and trust) to guide the design of model-driven visualizations. Contributed strategies (align, verify, modify, progressive disclosure) to aid designers in achieving interpretability and trustworthiness in visual analysis tools.
Abstract » Statistical topic models can help analysts discover patterns in large text corpora by identifying recurring sets of words and enabling exploration by topical concepts. However, understanding and validating the output of these models can itself be a challenging analysis task. In this paper, we offer two design considerations--interpretation and trust--for designing visualizations based on data-driven models. Interpretation refers to the facility with which an analyst makes inferences about the data through the lens of a model abstraction. Trust refers to the actual and perceived accuracy of an analyst's inferences. These considerations derive from our experiences developing the Stanford Dissertation Browser, a tool for exploring over 9,000 Ph.D. theses by topical similarity, and a subsequent review of existing literature. We contribute a novel similarity measure for text collections based on a notion of "word-borrowing" that arose from an iterative design process. Based on our experiences and a literature review, we distill a set of design recommendations and describe how they promote interpretable and trustworthy visual analysis tools.ACM
V-Model: A New Innovative Model to Chronologically Visualize Narrative Clinical Texts
Contribution & Benefit: Proposes and verifies an innovative timeline model for narrative clinical events. Solves natural language representation problems, provides information for temporal reasoning, and is intuitive for understanding patient histories.
Abstract » Visualizing narrative medical events into a timeline can have positive effects on clinical environments. However, the characteristics of natural language and medical environments make this representation more difficult. This paper explains the obstacles and suggests a solution called the V-Model. The V-Model is a new innovative time model that was developed to represent chronological narrative events in a medical domain. Forty medical students participated in evaluating this model. The experimental results show the new model successfully solved the modeling requirements and had better usability compared to conventional timeline models. All the participants assessed the new timeline as very useful in effectively understanding a patient's history.ACM
JigsawMap: Connecting the Past to the Future by Mapping Historical Textual Cadasters
Contribution & Benefit: We present an interactive visualization tool for visualizing and mapping historical textual cadasters. It can help historians understand the social/economic background of changes in land uses or ownership.
Abstract » In this paper, we present an interactive visualization tool, JigsawMap, for visualizing and mapping historical textual cadasters. A cadaster is an official register that records land properties (e.g., location, ownership, value and size) for land valuation and taxation. Such mapping of old and new cadasters can help historians understand the social/economic background of changes in land uses or ownership. With JigsawMap, historians can continue mapping older or newer cadasters. In this way, JigsawMap can connect the past land survey results to today and to the future. We conducted usability studies and long term case studies to evaluate JigsawMap, and received positive responses. As well as summarizing the evaluation results, we also present design guidelines for participatory design projects with historians.ACM
Semantic Interaction for Visual Text Analytics
Contribution & Benefit: Description of design space for user interaction for visual analytics called Semantic Interaction, coupling foraging and synthesis stages of sensemaking. The system, ForceSPIRE, supports users throughout sensemaking for text documents.
Abstract » Visual analytics emphasizes sensemaking of large, complex datasets through interactively exploring visualizations generated by statistical models. For example, dimensionality reduction methods use various similarity metrics to visualize textual document collections in a spatial metaphor, where similarities between documents are approximately represented through their relative spatial distances to each other in a 2D layout. This metaphor is designed to mimic analysts� mental models of the document collection and support their analytic processes, such as clustering similar documents together. However, in current methods, users must interact with such visualizations using controls external to the visual metaphor, such as sliders, menus, or text fields, to directly control underlying model parameters that they do not understand and that do not relate to their analytic process occurring within the visual metaphor. In this paper, we present the opportunity for a new design space for visual analytic interaction, called semantic interaction, which seeks to enable analysts to spatially interact with such models directly within the visual metaphor using interactions that derive from their analytic process, such as searching, highlighting, annotating, and repositioning documents. Further, we demonstrate how semantic interactions can be implemented using machine learning techniques in a visual analytic tool, called ForceSPIRE, for interactive analysis of textual data within a spatial visualization. Analysts can express their expert domain knowledge about the documents by simply moving them, which guides the underlying model to improve the overall layout, taking the user�s feedback into account.ACM