Frameworks for Information Visualization
Motivation “The purpose of visualization is insight, not pictures. The main goals of this insight are discovery, decision making, and explanation” Card, Mackinlay, Shneiderman, Reading in information visualization: using vision to think. 1999
Overview A Knowledge Task-Based Framework for Design and Evaluation of Information Visualization Robert Amar, John Stasko Distributed Cognition as a Theoretical Framework for Information Visualization Zhicheng liu, Nancy J. Nersessian, and John T. Stasko
Overview A Knowledge Task-Based Framework for Design and Evaluation of Information Visualization Robert Amar, John Stasko Distributed Cognition as a Theoretical Framework for Information Visualization Zhicheng liu, Nancy J. Nersessian, and John T. Stasko
Representational Primacy Definition: The pursuit of faithful data replication and comprehension Can be limiting Focus to much on low-level tasks that do not map well to the true needs of users
Goals Learning a domain Complex Decision making under uncertainty
Analytic Gaps Worldview Gap Rationale Gap Representation of Data Analyst Perceptual Processes Worldview Gap Perceiving Useful Relationships Worldview Gap: The gap between what is being shown and what actually needs to be shown to draw a straightforward represenational conclusion for making a decision Rationale Gap: The gap between perceiving a relationship and actually being able to explain confidence in that relationship and the usefulness of that relationship Rationale Gap Higher-Level Analytic Activity Explaining Relationships Robert Amar: InfoVis 2004
Worldview Gap: The Gap between what is being shown and what actually needs to be shown to draw a straightforward to draw a straightforward representational conclusion for making a decision Robert Amar: InfoVis 2004
Analytic Gaps Worldview Gap Rationale Gap Representation of Data Analyst Perceptual Processes Worldview Gap Perceiving Useful Relationships Worldview Gap: The gap between what is being shown and what actually needs to be shown to draw a straightforward represenational conclusion for making a decision Rationale Gap: The gap between perceiving a relationship and actually being able to explain confidence in that relationship and the usefulness of that relationship Rationale Gap Higher-Level Analytic Activity Explaining Relationships Robert Amar: InfoVis 2004
Rationale gap: The gap between perceiving a relationship and actually being able to explain confidence in that relationship and the usefulness of that relationship Robert Amar: InfoVis 2004
Analytic Gaps Worldview Gap Rationale Gap Representation of Data Analyst Perceptual Processes Worldview Gap Perceiving Useful Relationships Worldview Gap: The gap between what is being shown and what actually needs to be shown to draw a straightforward represenational conclusion for making a decision Rationale Gap: The gap between perceiving a relationship and actually being able to explain confidence in that relationship and the usefulness of that relationship Rationale Gap Higher-Level Analytic Activity Explaining Relationships Robert Amar: InfoVis 2004
Knowledge Tasks Worldview Tasks Rationale Tasks Domain Parameters Multivariate Explanation Confirm Hypotheses Rationale Tasks Expose Uncertainty Concretize Relationships Formulate Cause and Effect Domain Parameters: Facilitate acquisition and transfer of knowledge and/or metadata about domain parameters - Multivariate Explanation: Support the discovery of useful correlative models – especially those involving many variables Confirm Hypotheses: Provide facilities for users to formulate and confirm hypotheses about the data set Expose Uncertainty: Expose the sources and effects of uncertainty in data measures and aggregations - Concretize Relationships: Show the elements comprising relationships and translate into real-world outcomes - Formulate Cause and Effect: Clarify the source and nature of possible causations Robert Amar: InfoVis 2004
Worldview Task 1: Determine Domain Parameters Facilitate acquisition and transfer of knowledge and/or metadata about domain parameters Robert Amar: InfoVis 2004
Worldview Task 2: Multivariate Explanation Support the discovery of useful correlative models – especially those involving many variables Robert Amar: InfoVis 2004
Worldview Task 3: Confirm Hypotheses Provide facilities for users to formulate and confirm hypotheses about the data set Robert Amar: InfoVis 2004
Rationale Task 1: Expose Uncertainty Expose the sources and effects of uncertainty in data measures and aggregations Good to reexplain “rationale” concept Robert Amar: InfoVis 2004
Rationale Task 1: Expose Uncertainty See Nick’s contest entry (ThemeExplorer?) – explicitly dealing with missing abstracts Consider augmenting this slide with “improvements” SeeIT (Visible Decisions) Grocery Store Spending Survey Visualization, Augmented Robert Amar: InfoVis 2004
Rationale Task 2: Concretize Relationships Show the elements comprising relationships and translate into real-world outcomes Robert Amar: InfoVis 2004
Rationale Task 3: Formulate Cause and Effect Clarify the source and nature of possible causations Robert Amar: InfoVis 2004
Knowledge Tasks Worldview Tasks Rationale Tasks Domain Parameters Multivariate Explanation Confirm Hypotheses Rationale Tasks Expose Uncertainty Concretize Relationships Formulate Cause and Effect Domain Parameters: Facilitate acquisition and transfer of knowledge and/or metadata about domain parameters - Multivariate Explanation: Support the discovery of useful correlative models – especially those involving many variables Confirm Hypotheses: Provide facilities for users to formulate and confirm hypotheses about the data set Expose Uncertainty: Expose the sources and effects of uncertainty in data measures and aggregations - Concretize Relationships: Show the elements comprising relationships and translate into real-world outcomes - Formulate Cause and Effect: Clarify the source and nature of possible causations
Using the Tasks Generate new subtasks for a visualization to support or perform. Identify possible shortcomings in representation or data. Discover possible relationships to highlight or use as the basis for a visualization.
Overview A Knowledge Task-Based Framework for Design and Evaluation of Information Visualization Robert Amar, John Stasko Distributed Cognition as a Theoretical Framework for Information Visualization Zhicheng liu, Nancy J. Nersessian, and John T. Stasko
Representations External: distributed cognitive activity is directly observable Internal: not observable, but can identify where and when they are being processed by observing external representations.
External Representations
Distributed Representations Rule 1: Only one disk can be transferred at a time Rule 2: a disk can only be transferred to a pole on which it will be the largest Rule 3: only the largest disk on a pole can be transferred to another pole. [Zhang and Normal, 1993]
Interaction The ability to modify one’s environment to save on internal computation Liu, InfoVis ‘08
Interaction [Kirsh and Maglio, 1994]
Distributed Cognition Interaction is used to coordinate an external and internal representation, making the environment an extension of one’s self
Distributed Cognition (Single) 2314 x 184
Distributed Cognition (Single) 2314 x 184
Distributed Cognition (Group) Length
Evaluation The Whole is greater then the sum of its parts In situ observations and ethnographic approaches of cognitive system Empirical observations used for developing theories and taxonomies
Benefits to considering theories Descriptive: Identify key concepts and provide a conceptual framework Explanatory: rhetorically support explaining relationships and processes to support education and training Predictive: make predictions about performance in existing and new situations Prescriptive: provide guidelines and warnings for design Generative: facilitate creativity and discovery in future research