Search site...

5 Papers about high-dimensional data visualization

In this post I would like to share a small review about 1 article and 4 papers about how to manage visualization for high dimensional data.

.

1. A tour through the visualization zoo

The less academic article, and yet one of the best about visualization because it is elementary and provides a snapshot of visualization techniques from the obvious to the complex.

It began established the fact that our society is producing data at an astonishing rate, but require visualizations to help to understand the data. The main challenge is to create compelling and engaging visualizations that are appropriate for the data.

Some of the type of vis explained in the article are:

  • Time-series data: Index charts, stacked graphs, small multiples and horizon graphs.
  • Statistical distributions: Stem-and-leaf plots, Q-Q plots, scatter plot matrix and parallel coordinates.
  • Maps: Flow maps, choropleth maps, graduated symbol maps and cartograms.
  • Hierarchies: Node links diagram, adjacency diagram and enclosure diagrams.
  • Networks: force directed layouts, arc diagrams and matrix views.

All visualization share a common “DNA” – a set of mappings between data properties and visual attributes such as position, size, shape and color- and customized species of visualization might always be created by varying the encodings.

Take away

  • A simple explanations about all the posible visualizations that you need to create visualization for high dimensional data.

Keywords

  • Time-series data | Index charts | stacked graphs | small multiples and horizon graphs | Statistical distributions | Stem-and-leaf plots | Q-Q plots | scatter plot matrix | parallel coordinates | Maps | Flow maps, choropleth maps | graduated symbol maps | cartograms | Hierarchies | Node links diagram | adjacency diagram | enclosure diagrams | Networks | force directed layouts | arc diagrams | matrix views

.

2. High dimensional data visualization by Martin Theus

This short paper made over the four different plot types has proven to be the most powerful at the moment to create visualizations for the multivariate structure of several variables at a time.

In general, we can distinguish two types of data displays: dynamic (i.e., which allow user interaction) and static. It turns out that in high dimensional data exploration the interaction with the visual display is very important to gain a better insight into the data. For presentation purposes, static graphics are usually needed to communicate the findings.

The plot types explained are:
• For purely categorical data: Mosaic Plots
• For purely continuous data: Parallel Coordinate Plots and Projection Pursuit and Grand Tour
• For data on mixed scales: Trellis Displays

Take away

  • A comprehensive explanation about 4 plots.

Keywords

  • Mosaic Plots | Parallel Coordinate Plots | Projection Pursuit and Grand Tour | Trellis Displays

.

3. High dimensional data visualization by Martin Theus from book Hand book of Data visualization

One of the biggest challenges in data visualization is to find general representations of data that can display the multivariate structure of more than two variables. Several graphic types like mosaic plots, parallel coordinate plots, trellis displays, and the grand tour have been developed over the last three decades. Each of these plots corresponds to a specific section of this handbook. This chapter will concentrate on investigating the strengths and weaknesses of these plots and techniques and contrast them in the light of data analysis problems. One significant issue is the aspect of interactivity. Except for trellis displays, all the above plots need interactive features to rise to their full power. Some, like the grand tour, is only defined by using dynamic graphics.

.

Take away

  • A solid and yet not excessively large academic chapter from a book of the reference in the field, where each of the visualizations are explained using examples and plots to illustrate all the concepts.

.

Keywords

  • Conditioning Variable | Common Scale | Projection Pursuit | Brand Person | Scaling Option | Mosaic plots | Trellis displays | Parallel coordinate plots | Grand Tour

.

4. Visualizing High-Dimensional Data: Advances in the Past Decade (2015)

.
Even when it is not the state of the art of academic publications related to high dimensional data, the article is not entirely outdated and provide a comprehensive survey of advances in high-dimensional data visualization over the past 15 years.

The paper explains that during the past decade, a variety of approaches have been introduced to visually convey high-dimensional structural information by utilizing low-dimensional projections or abstractions: from dimension reduction to visual encoding, and from quantitative analysis to interactive exploration.

Also, it points out different surveys oriented to various aspects of high-dimensional data visualization, such as parallel coordinates, quality measures, clutter reduction, visual data mining, and interactive techniques.

They propose a categorization of recent advances based on the information visualization (InfoVis) pipeline.

.

The authors highlight the key contributions of each advancement. Also, they connect advances in high-dimensional data visualization with volume rendering and machine learning.

Take away

  • Categorization of recent advances based on the information visualization
  • Highlights of the key contributions of each advancement
  • Connection between advances and machine learning.
  • References between a variety of papers refer to each particular area of concern.

.

Keywords

  • Categorization | information visualization pipeline | survey of visualization techniques | metrics | high-dimensional data visualization | visualization pipeline | Data models | Taxonomy | high-dimensional data | multidimensional data

.

5. High-Dimensional Visualizations

In this paper the authors provide a brief background to data visualization
and point to key references. They differentiate between high dimensional data visualization and high-dimensional data
visualizations and review the various high-dimensional
visualization techniques.

Another remarkable point is the variety of visualization included to facilitate the process of learn and understand different concepts.

.

Take aways

  • Plots accompanying each of the concepts explained.
  • Simple and plain language to explain concepts related with data visualization
  • A summary of intrinsic properties for visualizations discussed

.

Keywords

  • visualization techniques overview | evaluation | high dimensional data visualization

.

Citations

  • A tour through the visualization zoo, Published by ACM 2010 Article. Popular; Refereed. Bibliometrics Data Bibliometrics
  • Theus M. (2008) High-dimensional Data Visualization. In: Handbook of Data Visualization. Springer Handbooks Comp.Statistics. Springer, Berlin, Heidelberg
  • Grinstein, Georges & Trutschl, Marjan. (2001). High-Dimensional Visualizations. In: 7th Data Mining Conference-KDD 2001: San Francisco, California.
  • S. Liu, D. Maljovec, B. Wang, P. Bremer and V. Pascucci, “Visualizing High-Dimensional Data: Advances in the Past Decade,” in IEEE Transactions on Visualization and Computer Graphics, vol. 23, no. 3, pp. 1249-1268, 1 March 2017.
    doi: 10.1109/TVCG.2016.2640960

Graphics | Table:

  • Each table or visualization used in this article belongs to the original authors of each paper, and it is only included to add information of each description.