Data Visualization

2.8

for data exploration. MU- May 17,Dec. 18

Q. ExDlain types of attributes and data visualization
visualization. MU-Dec17
Q. Write short note on Data
people
Data visualisation is presenting the data in a graphical or pictorial format. Visualisation techniques help
the data can be marked very
to analyse things which are otherwise not possible when the data is large. Patterns in
easily using the data visualisation techniques. Some of the data visualisation techniques are as follows:

Tech kaesiedge
 Intro. to Data Mining, Data
Data Warehousing and Mining (MU) 2-20
Exploration and Data
1. Pixel-orlented visualization techniques PTe-r
In pixel based visualisation technigues, there is a separate subwindows for the value of each
represented by one colored pixel.
overlap.
attrìbute and
lt maximises the amount of information represented at one time without any
each variable and each
Atuple with m variables has different m colored pixel to represent
window. yariable has s
mapping of the pixels is decided.
Based on data characteristics and visualization task, the color
A
dala Record with
four variables

Fig. 2.8.1: Pixel visualisation with

four variables

2. Geometric projection visualization techniques hrotbsesste

sets can be found using the
Geometric transformations and projections of multidimensional data
techniques :
(a) Scatterplot matrices
(b) Hyper slice
(c) Parallel coordinates

variables in a dataset.
(a) Scatterplot matrices: It is composed of scatter plots of all possible pairs of
Climato predictors

WetDays

TempJuly

TempJan

TernpAnn

RHuly
otesilsualy si
Fig. 2.8.2:An example of scatter plot
(b) Hyperslice : It is an extension to scatter plot matrices. They represent a multidimensional function a
matrix oforthogonal two dimensional slices.
(c) Parallel co-ordinates: The parallel vertical lines separated define the axes. A point in the Carte
coordinates corresponds to a polyline in parallel coordinates.

TechKsee
PuDiicath
 Warehousing and|Mining (MU)
Data 2-21 Intro. to Data Mining, Data Exploration and Data Pre-pro.
4.5
25

oelsdode no 36

2.5

Sepal width Sepal length Petal width Petal length

Setosa-Versicolor Virginica
Fig. 2.8.3 : An example of Parallel
coordinates
lcon-based visualization techniques
Ieon based visualisation techniques also known as iconic
display techniques.
Bach multidimensional data item is mapped to an icon. 6apiioet nolfosilasely laoloncot
This technique allows visualisation of large amountsof fdata.
daralh beennalbgpiacod tcdali
Two most commonly used icon based techniques are: e aotihass 2ouptedbs totseleuvloti
(a) Chernoff faces
(b) Stick figures
gnbbste isnolznonil
(a) Chernoff faces (
" Illustration of trends in multidimensional data can be done by using Chernoff faces. This concept was
introduced by Herman Chernoff in the year 1973.
" The faces in Chernoff faces are related to facial expressions or features of human being. So to distinguish
between them is easy.
" Different data dimensions were mapped to different facial features, for example the face width, the length or
curvature of the mouth,the length of the nose etc.
" An example of Chernoff faces is shown in Fig. 2.8.4; they use facial features to represent trends in the values of
the data, not the specific values themselves. They display multidimensional data of upto 18 variables or
dimensions. In Fig, 2.8.4, each face represents an n-dimensional data points (n<=18).

ehidetiA

Fig. 2.8.4: An example of Chernoff faces

Techknewledyd
 Data Warehousing and Mining (MU) 2-22 Intro. to Data Mining, Data Exploration and Data
(b) Stick figures Pre-prO.
Pickett and Grinstein introduced stick figure icon visualisation technique.
The Fig. 2.8.5 represents the original stick figure with five stick and family of twelve of them.
This icon family is designed to display data with up to five variates.
Stick icon can be used to display bivariate MRIdata by using a two stick icon which helps to differentiate
texture in a complex image.

(a) Afive stick figure (b) Astick figure icon family

icon with orientation with a body and four limb
pliast noltssitnoiy b
Fig. 2.8.5:Example of stick figure
4. Hierarchical visualization techniques
The visualisation techniques discussed above display multiple dimensions simultaneously. However for a large
data set having large number of dimensions the above techniques may not be useful.
Hierarchical visualisation techniques partition alldimensions in to subset (subspaces).oaenesaon
These subspaces are visualised in ahierarchical manner.
Some of the visualisation techniques are :
(a) Dimensional stacking
(b) Mosaic Plot
(c) Worlds-within-worlds
(d) Tree-map
(e) Visualizing complex data and relations

(a) Dimensional stacking

In dímension stackíng partítion the n-dimensional attribute space in 2-dimenstinal subspaces.
Attribute values are partitioned into various classes.
Each element is a two dimensional space is a xy plot.
Mark the important attributes and are used on the outer levels.

Attribute 4
Attribute 2

Attribute 3

Attribute 1

Fig. 2.8.6 : Data in dimension stacking

Tech Kouledga
PubICa1
 Data Warehousing and Mining (MU)
2-23 Intro. to Data Mining, Data Exploration and Data Pre-pra
(b) Mosaic plot
Heeaic plots give a graphical
illustration of the successive
decompositions.
. Rectangles are used to represent the count of
To draw a mosaic plot, a categorical data and at every stage rectangles are split
paraleL
is required.
contingency table of data and chosen ordering of variable
with the response variable
Example: In titanic example, out of all women, 67% survived
as0. So the women bar shows as 67/33 split. which is coded as 1 and 33% died
Among men, only 17% survived, so this bar shows a which coded
is
Women Men
17/83 split.
Survived

Sex
Fig. 2.8.7: Mosaic Plot for Titanic
(c) Worlds within worlds
" Worlds within worlds are
useful to generate an interactive hierarchy of display.
" Innermost world mustt have a function and two most
important parameters.
" Remaining parameters fix with constant value.
Through this N-vision of data are possible like data glove and
(inner) and translation stereo displays, including rotation, scaling
(inner/outer)dlsdnly
Using queries static interaction is also possible.
sisoe

Fig. 2.8.8: Worlds within worlds visualization

TechKrouledgi
PbIC atian.
 Intro. to Data Mining, Data
DataWarehousing and Mining (MU) 2-24
Exploration and Data
(d) Tree-maps
large amounts of
Pre-pro,
Tree maps visualization techniques are well suited for displaying hierarchical
data.

The visualization space is divided into multiple rectangles that are sized and ordered
structured
quantitative variable. according
rectangles.
The levels in the hierarchy are seen rectangles containing other
Each set of rectangles on the same level in the hierarchy represents a column or an expression in a data set.

Glabal

Earope
Ametcas

rael
Spait
nited Kngdot

etternds

Iretand
Russia
Germaiy.
Poland

Me.
Austrála
Aka

Fig. 2.8.9 : Web traffic by location Tree-map

category in a column.
Each individual rectangle on a level in the hierarchy represents a
which there are rectangles representing
For example, in the Fig. 2.8.9, a rectangle representing global below
that continent.a
continents which contain several rectangles representing countries in
representing states in these countries.
Each rectangle representing a country may in turn contain rectangles
(e) Visualizing complex data and relations
product reviewS.
Thís technique is useful to visualize non-nuneric data such as text, pictures, blog entries and
" A tag cloud is a visualization method which helps to understand the information of user generated tags.
and colors.
Arrange the tags alphabetically or with the user preferences with different font sizes
that how many times that tag
Tag clouds are used in two ways that with the size of tag, we can find out
applied on that item by different users or that tag has been applied to how many items.

design tesk h A e

information
herUSe Teaiuresoom
interlace.team
Users k o ime.
torsopeonsUSabilitynay
ess cemets ,Oneetra
dleret Pprodct data
ing
also
application
Fig. 2.8.10 : Social data visualisation TechKnol
Peott
catt2#