Information visualization fundaments

Definition (chapter Introduction and fundaments)

Visual analytics combines automated analysis techniques with

interactive visualizations for an effective understanding,

reasoning and decision making on the basis of very large and

complex datasets

[Keim]

Visual analytics process

Knowledge

Models

Model building Modelvisualization

Visual data exploration

Automated data analysisFeedback loop

Datamining

Mapping

User interaction

Preprocessing and transformation

Parameterrefinement

Dataset

Visualization

Visual analytics mantra

Analyze first, show the important, zoom/filter, analyze further, details on demand

D. A. Keim, F. Mansmann, J. Schneidewind, and H. Ziegler. Challenges in visual data analysis. In Information Visualization (IV 2006), Invited Paper, July 5-7, London, United Kingdom. IEEE Press, 2006.

Multidisciplinary approach

Visualization

Human perception and cognition

Data mining

Data management

Spatio-temporal data analysis

Infrastructures

Evaluation

Contents

Visualization need Advantages Design issues Abstraction levels Validation

Definition

Need

Computer-based visualization systems provide visual representations of datasets designed to help people carry out tasks more effectively

[Munzner]

Visualization is suitable when there is a need to augment human capabilities rather than replace people with computational decision-making methods

[Munzner]

Need

Visualizations help people analyzing data when they don't know exactly what questions they need to ask in advance

External representations implemented in computers improve human capacity and allow to surpass the limitations of our internal cognition and memory.

Advantages of the visual channel

Human visual system is a high bandwidth channel that provides to our brains an enormous quantity of information. It can be processed in parallel at the pre-conscious level or in a conscious way focusing our attention over visible objects.

The input of the audio channel is perceived as a sequential stream of sounds instead of gather them during some period of time for being processed altogether.

Interaction with other senses still hampers from technological issues that prevent its use outside from the research sphere.

Advantages of the visual channel

Visualization tools help human beings in those situations where seeing data structure is better than obtaining a brief summary. For example: The Anscombe quartet.

Interactivity is a key feature for building visualization tools that manage complex dataset, either because they are large enough or because they vary in time.

Task abstractions

In the process of designing visualization tools, user tasks are as crucial as the dataset.

At the higher level of abstraction, we can distinguish four categories of user tasks: presentation, discovery, enjoyment and production of new information.

Design considerations

Effectiveness is a corollary of defining visualization to have the goal of supporting user tasks.

Instead of searching the optimal solution, a more adequate goal designing visualization tools is to satisfy the requirements.

Design considerations

We can structure the design process answering the following questions:

Why users intend to use a visualization tool? What data users see? How the visual encoding and user interactions are implemented

from the point of view of design choices? The procedure of answering these questions can be iteratively applied

until we obtain the final design solution.

Design abstraction levelsProblem characterization in the application domain

Data and task abstractions

Interaction and visual encoding

Algorithmic implementation

Design abstraction levels

These levels are nested: the output from an upstream level above is input to the downstream level below. A block is the outcome of the design process at that level, and must be validated before being used in the next design level.

User intervention is critical for validating the designs in the three higher levels of abstraction.

Focus design methodology through user needs is called user-centered design.

Problem characterization

In this abstraction level we describe specific issues of the application domain and end users involved, such as the problem to solve, user demands and datasets.

Each application domain has its own argot and its own workflow. Designers obtain outcomes at this level through end users interviews,

through direct observation of their work in real work conditions and researching about end users work.

Data and task abstractions

We must make abstractions of the specific tasks and data involved in the application domain and map them to generic representations independent from the concrete application domain.

For tasks, we must identify the tasks required by end users in their workflow.

For example: explore, compare, resume. For data, the goal will be to determine which is the representation

that best fits users needs.

Interaction and visual encoding

At this level we must determine the specific design choice for creating and manipulating the visual representations of the abstract data types that have been selected in the upper abstraction level, guided by the abstract tasks identified at that level.

For visual encoding, designers define what users see. For interaction, designers decide how are dynamically managed data

representations, i. e., how users change what they are seeing.

Algorithmic implementation

In this case, the goal is to achieve an efficient implementation of the visual encoding and the interaction techniques selected in the previous abstraction level.

The selection of one or another algorithm will depend on the specific requirements of the problem and the available resources.

Design abstraction levels: validationRisk: wrong problem characterizationValidation: observe and interview end users

Validation: measure adoption

Risk: wrong data and task abstractions

Validation: end user tests, collect proofs about the system utilityValidation: field studies, usage of the deployed system

Risk: inefficient interaction and visual encodingValidation: justify the visual encoding and the interaction techniques selected

Validation: quantitative and qualitative analysis of user feedback (user study)Validation: laboratory studies, measure of user errors and user time operation

Risk: slow algorithmValidation: algorithm complexity analysis

Validation: measure system response time and memory used

System implementation

Design approaches

Top-down: problem-driven. Bottom-up: technique-driven.

References

[Munzner]: Tamara Munzner. Visualization Analysis and Design. A K Peters Visualization Series. CRC Press. Nov. 2014.

[Anscombe]: F.J. Anscombe. Graphs in Statistical Analysis. American Statistician 27 (1973), 17-21.

PrototypeimplementationDesign

ValidationAnalysis

Abstract tasks Abstract data and views

MethodsHow?

What?Why?

How?

What?Why?

How?

What?Why?

How?

What?Why?

Information visualization fundamentsDefinition (chapter Introduction and fundaments)Visual analytics processVisual analytics mantraMultidisciplinary approachContentsDefinitionNeedAdvantages of the visual channelAdvantages of the visual channelTask abstractionsDesign considerationsDesign considerationsDesign abstraction levelsDesign abstraction levelsProblem characterizationData and task abstractionsInteraction and visual encodingAlgorithmic implementationDesign abstraction levels: validationDesign approachesReferencesSlide Number 23Slide Number 24Slide Number 25Slide Number 26