Multidimensional Projection for Visual Analytics: Linking Techniques with Distortions, Tasks, and Layout Enrichment

Luis Gustavo Nonato, Michael Aupetit

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Visual analysis of multidimensional data requires effective ways to reduce data dimensionality to encode them visually. Multidimensional projections (MDP) figure among the most important visualization techniques in this context, transforming multidimensional data into scatter plots where patterns reflect some notion of similarity in the data. However, MDP come with distortions that make visual patterns not trustworthy. Moreover, the patterns present in scatter plots might not be enough to allow an understanding of multidimensional data, motivating the development of layout enrichment methodologies that operate with MDP. This survey attempts to cover the main aspects of MDP as a visualization and visual analytic tool, providing detailed analysis and taxonomies taht organize MDP techniques according to their main properties and traits. The survey also approaches the different types of distortions that can result from MDP mappings while overviewing existing mechanisms to quantitatively evaluate such distortions. A qualitative analysis of the impact of distortions on the different analytic tasks is also presented, providing guidelines for users to choose a proper MDP for an intended. Finally, layout enrichment schemes to debunk MDP distortions and/or reveal relevant information not directly inferable from the scatter plot are reviewed and discussed in the light of new taxonomies.

Original languageEnglish
JournalIEEE Transactions on Visualization and Computer Graphics
DOIs
Publication statusAccepted/In press - 12 Jun 2018

Fingerprint

Taxonomies
Guidelines
Visualization
Surveys and Questionnaires

Keywords

  • Data visualization
  • Dimensionality Reduction
  • Distortion
  • Error Analysis
  • Layout
  • Layout Enrichment
  • Multidimensional Projection
  • Multidimensional Scaling
  • Task analysis
  • Taxonomy
  • Visual perception
  • Visualization

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Computer Graphics and Computer-Aided Design

Cite this

@article{0dfd782cdaf84420a30638d8cd34a46c,
title = "Multidimensional Projection for Visual Analytics: Linking Techniques with Distortions, Tasks, and Layout Enrichment",
abstract = "Visual analysis of multidimensional data requires effective ways to reduce data dimensionality to encode them visually. Multidimensional projections (MDP) figure among the most important visualization techniques in this context, transforming multidimensional data into scatter plots where patterns reflect some notion of similarity in the data. However, MDP come with distortions that make visual patterns not trustworthy. Moreover, the patterns present in scatter plots might not be enough to allow an understanding of multidimensional data, motivating the development of layout enrichment methodologies that operate with MDP. This survey attempts to cover the main aspects of MDP as a visualization and visual analytic tool, providing detailed analysis and taxonomies taht organize MDP techniques according to their main properties and traits. The survey also approaches the different types of distortions that can result from MDP mappings while overviewing existing mechanisms to quantitatively evaluate such distortions. A qualitative analysis of the impact of distortions on the different analytic tasks is also presented, providing guidelines for users to choose a proper MDP for an intended. Finally, layout enrichment schemes to debunk MDP distortions and/or reveal relevant information not directly inferable from the scatter plot are reviewed and discussed in the light of new taxonomies.",
keywords = "Data visualization, Dimensionality Reduction, Distortion, Error Analysis, Layout, Layout Enrichment, Multidimensional Projection, Multidimensional Scaling, Task analysis, Taxonomy, Visual perception, Visualization",
author = "Nonato, {Luis Gustavo} and Michael Aupetit",
year = "2018",
month = "6",
day = "12",
doi = "10.1109/TVCG.2018.2846735",
language = "English",
journal = "IEEE Transactions on Visualization and Computer Graphics",
issn = "1077-2626",
publisher = "IEEE Computer Society",

}

TY - JOUR

T1 - Multidimensional Projection for Visual Analytics

T2 - Linking Techniques with Distortions, Tasks, and Layout Enrichment

AU - Nonato, Luis Gustavo

AU - Aupetit, Michael

PY - 2018/6/12

Y1 - 2018/6/12

N2 - Visual analysis of multidimensional data requires effective ways to reduce data dimensionality to encode them visually. Multidimensional projections (MDP) figure among the most important visualization techniques in this context, transforming multidimensional data into scatter plots where patterns reflect some notion of similarity in the data. However, MDP come with distortions that make visual patterns not trustworthy. Moreover, the patterns present in scatter plots might not be enough to allow an understanding of multidimensional data, motivating the development of layout enrichment methodologies that operate with MDP. This survey attempts to cover the main aspects of MDP as a visualization and visual analytic tool, providing detailed analysis and taxonomies taht organize MDP techniques according to their main properties and traits. The survey also approaches the different types of distortions that can result from MDP mappings while overviewing existing mechanisms to quantitatively evaluate such distortions. A qualitative analysis of the impact of distortions on the different analytic tasks is also presented, providing guidelines for users to choose a proper MDP for an intended. Finally, layout enrichment schemes to debunk MDP distortions and/or reveal relevant information not directly inferable from the scatter plot are reviewed and discussed in the light of new taxonomies.

AB - Visual analysis of multidimensional data requires effective ways to reduce data dimensionality to encode them visually. Multidimensional projections (MDP) figure among the most important visualization techniques in this context, transforming multidimensional data into scatter plots where patterns reflect some notion of similarity in the data. However, MDP come with distortions that make visual patterns not trustworthy. Moreover, the patterns present in scatter plots might not be enough to allow an understanding of multidimensional data, motivating the development of layout enrichment methodologies that operate with MDP. This survey attempts to cover the main aspects of MDP as a visualization and visual analytic tool, providing detailed analysis and taxonomies taht organize MDP techniques according to their main properties and traits. The survey also approaches the different types of distortions that can result from MDP mappings while overviewing existing mechanisms to quantitatively evaluate such distortions. A qualitative analysis of the impact of distortions on the different analytic tasks is also presented, providing guidelines for users to choose a proper MDP for an intended. Finally, layout enrichment schemes to debunk MDP distortions and/or reveal relevant information not directly inferable from the scatter plot are reviewed and discussed in the light of new taxonomies.

KW - Data visualization

KW - Dimensionality Reduction

KW - Distortion

KW - Error Analysis

KW - Layout

KW - Layout Enrichment

KW - Multidimensional Projection

KW - Multidimensional Scaling

KW - Task analysis

KW - Taxonomy

KW - Visual perception

KW - Visualization

UR - http://www.scopus.com/inward/record.url?scp=85048564304&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048564304&partnerID=8YFLogxK

U2 - 10.1109/TVCG.2018.2846735

DO - 10.1109/TVCG.2018.2846735

M3 - Article

AN - SCOPUS:85048564304

JO - IEEE Transactions on Visualization and Computer Graphics

JF - IEEE Transactions on Visualization and Computer Graphics

SN - 1077-2626

ER -