Scribd
Upload
30 views

Informatics: Lecture 7 - Visualising Information

This document discusses visualizing data and information visualization. It provides definitions and examples of visualizing different types of data, from univariate to multivariate. The key points are: 1) Visualizing data can help discover patterns, analyze relationships, and illustrate abstract concepts. It allows people to make sense of complex data and findings more easily. 2) Information visualization uses visual elements like color, size, shape to communicate relationships and illustrate information. 3) Different types of visualizations are suited for different dimensions of data, from histograms for univariate, scatter plots for bivariate, to using multiple visual parameters for hypervariate/multidimensional data. Examples like spider diagrams, mosaic plots, and spatial encoding are provided.

Uploaded by

Colin
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
30 views

Informatics: Lecture 7 - Visualising Information

This document discusses visualizing data and information visualization. It provides definitions and examples of visualizing different types of data, from univariate to multivariate. The key points are: 1) Visualizing data can help discover patterns, analyze relationships, and illustrate abstract concepts. It allows people to make sense of complex data and findings more easily. 2) Information visualization uses visual elements like color, size, shape to communicate relationships and illustrate information. 3) Different types of visualizations are suited for different dimensions of data, from histograms for univariate, scatter plots for bivariate, to using multiple visual parameters for hypervariate/multidimensional data. Examples like spider diagrams, mosaic plots, and spatial encoding are provided.

Uploaded by

Colin
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 33

Informatics

Lecture 7 Visualising Information

Sources
Information Visualization Spence
Visualizing Data Fry
Information is Beautiful McCandless

Why bother to visualise data?


Why do we bother with trying to
visualise data sets why dont we
just mine the data with some cool
techniques and just get the answer
that we need?

Answer 1
There may not be a simple answer
and we need to make some sort of
judgment
People are very good at making
sense of noisy data and finding
meaning in complex data.

Answer 2
We may need to present our
findings to persuade others to
share our conclusion or point of
view information visualisation for
advocacy

Answer
A picture may summarise a whole
series of events or connections
between things and these results
dont have an answer as such

A definition
Information visualisation =
Information Design
..uses pictures, symbols, colours
and words to communicate ideas,
illustrate information or express
relationships visually.or aurally

Is this effective?

Applications of Information Design


Persuade people advocacy
Discover patterns and trends
Analyse relationships/connections
Evaluate factors
Evaluate scenarios
Illustrate the abstract
Highlight a trend or change

Nature of the problem


We will often be faced with
incomplete, inaccurate or
contradictory data. The data will
often be multi-dimensional which
presents challenges as to how to
illustrate it.

Example choosing a car


Make

Price

mpg

Rating

Age (yrs)

Ford

15,540

31

*****

GM

12,450

27

***

etc

The data is 5-dimensional:


Make category
Price numerical
mpg numerical
Rating ordinal
Age numerical
How can we view this to help someone
to make a choice?

Representation issues
We can see that even for this simple
example the data is complex and
multi-dimensional or hypervariate
Perhaps we could use a simple
scatter plot?

GOOD

mpg

BAD
Price

So what about rating and age?

Hypervariate data
We can use scatter plots for 2D data
but for more dimensions we need to
introduce other visual parameters
Lets work up through the
dimensions.

Univariate data 1D

We have a single value to consider


but even here good choices of the
display method can save lives

Univariate data
Even for this simple case we need to
consider:
Is it important to notice if the
single value changes?
What if the value is not accurate?
Of course the histogram is a
favourite way to display this data

Bivariate data
With 2 dimensions, a scatter plot is
the most common choice.you may
recognise the following example

Trivariate data
It would seem natural to use 3D
plots for trivariate data but it can be
hard to compare data depending on
the viewing perspective 3D data
should be presented with controls to
vary the viewing perspective

Hypervariate data
With multi-dimensional data we
often use 2D plots (because of the
problem with 3D plots) and then use
other visual aspects to represent the
other dimensions.
There are a lot of potential
candidates for these other variables

Hypervariate data
Size area or a linear dimension?
Colour
Orientation
Symbol
Shape
Pattern
Plus others and combinations

Example spider diagram

Example mosaic plot

Example spatial encoding

You might also like