Visualization and Visual Analytics for Data Science

J. J. van Wijk | Saturday, July 13
Full professor in visualization at the Department of Mathematics and Computer Science of Eindhoven University of Technology (TU/e)

Visualization is a powerful method to explore, analyze, and present
data. Exploiting the unique capabilities of the human visual system
enables us to bring the human in the loop, to inspect the data, to
detect unexpected patterns, and to reason about case at hand. However,
visualization does have limitations. Human perception and cognition have
limitations and biases, which, if not taken into account, can lead to
confusing or misleading results; large data sets often contain too much
information to be grasped. Integration of methods from statistics,
machine learning, and data mining is an powerful way to deal with large
data, which is known as visual analytics. In the course, we study
visualization and visual analytics, discussing opportunities,
limitations, and challenges. Some exercises will be unplugged, make sure
that you bring pens and paper with you!

Overview of the lecture

Part 1: Visualization

Why visualization? An overview of the field is given, with examples for
the visualization of multivariate data, hierarchical data, and network
data.

Part 2: Perception

Understanding human perception is key in designing effective
visualization solutions. We discuss the strengths and weaknesses of
different visual channels, and how these can be optimally used.

Part 3: Interaction

Interaction enables us to explore much larger data-sets and get a better
understanding. We discuss different modalities and approaches for
interaction, including VR, multiple views, and exploration.

Part 4: Tools

A short overview is given of various ways to implement visualization.
There is no ideal solution, sophistication and simplicity have to be
balanced.

Part 5: Visual Analytics

Integration of automated methods enables us to explore even larger
data-sets, aiming to exploit the strengths of humans and machines, which
will be illustrated by examples. But also, models from machine learning
itself provide challenges: how to show how automated decisions are made?

Textbooks

Tamara Munzner. /Visualization Analysis and Design/. Taylor & Francis
(2014).

Colin Ware. /Information Visualization – Perception for Design/. 3rd
edition. Elsevier Science & Technology (2012).