DAT A VIS U ALIZ AT IO N

CO M M U N ICAT IN G W IT H DAT A

P R AM O D K U M AR N AIK
CH AIR P E R S O N AI AN D R O B O T ICS
DS U B AN G ALO R E
DAT A VIS U ALIZ AT IO N
What are the problems with this graph?
Display the data table

Calories in Common Candies

250
200
150
100
50
0

fy
Taf
i nt Corn m ar s i s t s . . . . l i ces r Balls
nerM andy g Gu y Be e
. .
Tw ocola. cola. cola. ectinS Sou
in C i n m i c o o P
f t erD C hew Gum Licor ilk Ch ilkCh ilkCh
A M M M

CO LU M N
CH AR T
ALT E R N AT E DIS P L AY

Sorting and expanding the scale of the graph allows all labels to be seen
as well as displaying a characteristic of the data.

Calories in Common Candies

250

200
150
100

50
0
y
pop ffee els Taff
olli lls ts To s n s
Gum rscotch L u r Ba i n ars ts Sli c e t or
C Car a m
Be a n rittle
wing Butte
S o
ligh
t M
y Be iceTwisPectin rMiCn andy elly nutB r Bar
Star m r i nn e J Pe a teBa late
Che Gum L i co
Afte r D s o c o l a
aorc o
Chip
s s Ball kCh nlkdCBh ts
late R aisin d Milk Dar A lmMoi anu
C ho c o
v e r e d
Ma l t e
c o l a te
e r e dPe
et o e Cho Cov
iSwe teC olat late
Sem cola hoc Milk oco
Cho ilkC h

CO LU M N
i l k M C
M Milk

CH AR T
O B J E CT IVE S
As you create graphics keep the following in mind.
 Avoid distortion of the true story.
 Induce the viewer to think about the substance, not the graph.
 Reveal the data at several layers of detail.
 Encourage the eye to compare different pieces.
 Support the statistical and verbal descriptions of the data.

STA6166-2-5
 DAT A VIS U ALIZ AT IO N Graphical Display
Objectives
The visual portrayal of quantitative information • Tabulation
Are used to: • Description
• Illustration
• Display the actual data table
• Exploration
• Display quantities derived from the
data
• Show what has been learned
about the data from other analyses
• Allow one to see what may be
occurring in the data over and
above what has already been
described

“A picture is w orth a thousand w ord s… ”

DAT A VIS U ALIZ AT IO N
W hat is d ata visualization?
“A picture speaks a thousand w ord s.” S imilarly an
infog raphic/visual can help us analyze d ata and hid d en
patterns in a much easier w ay.
IN T R O DU CT IO N T O DAT A VIS U ALIZ AT IO N
IN T R O DU CT IO N T O DAT A VIS U ALIZ AT IO N
W hy visualize d ata?

Data visualization is a w ay you can create a story throug h your d ata. W hen d ata is complex and und erstand ing
the micro-d etails is essential,the b est w ay is to analyze d ata throug h visuals.

Visuals can b e used for tw o purposes:

1 . E xploratory d ata analysis: T his is used b y d ata analysts,statisticians,and d ata scientists to b etter
und erstand d ata. As it is rig htly called ,it is used to explore the hid d en trend s,patterns in d ata.

2 . E xplanatory d ata analysis: O nce the analysts und erstand the d ata and find their results,the b est w ay to
convey their id eas and find ing s is throug h visuals! T his is used to craft a story that w ill appeal to the view er
offering d eeper insig hts.
Importance of P lotting
The purpose of plotting scientific data is to visualize variation

or show relationships between variables, but not all data sets

require a plot. If there are only one or two points, it is easy to

examine the numbers directly, and little or nothing is gained by

putting them on a graph. Similarly, if there is no variation in

the data, it is easy enough to see or state the fact without using

a graph of any sort.

Line P lot

Line plots is a graph that is used for the representation of

continuous data points on a number line. Line plots are

created by first plotting data points on the Cartesian plane

then joining those points with a number line. Line plots

can help display data points for both single variable

analysis as well as multiple variable analysis.

S catter P lot

A scatter plot (also called a scatterplot, scatter g raph,

scatter chart, scatterg ram, or scatter d iag ram) is a type of

plot or mathematical d iag ram using Cartesian coord inates

to d isplay values for typically tw o variab les for a set of

d ata.
S catter P lot for M achine Learning
T hese are the charts/plots that are used to ob serve
and d isplay relationships b etw een variab les using
Cartesian Coord inates. T he values (x: first variab le ,y:
second variab le) of the variab les are represented b y
d ots. S catter plots are also know n as scatterg rams,
scatter g raphs, scatter charts ,or scatter d iag rams. It
is b est suited for situations w here the d epend ent
variab le can have multiple values for the ind epend ent
variab le.
B ar Chart

B ar charts are b est suited for the visualization of

categ orical d ata b ecause they allow you to easily see the

d ifference b etw een feature values b y measuring the

size(leng th) of the b ars. T here are 2 types of b ar charts

d epend ing upon their orientation (i.e. vertical or

horizontal).
G rouped B ar Chart
B ox P lot

B ox-plots tell us ab out the d istrib ution of d ata and scan

for outliers.

Also notice that even thoug h the numb er of nod es is a

more useful feature,there is some overlap w ith b oth the
classes.
Violin-P lot

A violin plot is a hyb rid of a b ox plot and a kernel d ensity plot,

w hich show s peaks in the d ata. It is used to visualize the
d istrib ution of numerical d ata.

U nlike a b ox plot that can only show summary statistics,

violin plots d epict summary statistics and the d ensity of each
variab le.
A heat map is a tw o-d imensional
representation of d ata in w hich values are
represented b y colors. A simple heat map
provid es an immed iate visual summary of
information. M ore elab orate heat maps
allow the view er to und erstand complex
d ata sets.
H eatmap

H eatmaps are the g raphical representation of

d ata w here each value is represented in a matrix
w ith d ifferent color cod ing . M ostly heatmaps are
used to find correlations b etw een various d ata
columns in a d ataset.
Speaker Details
Email:[email protected]
[email protected]
Mobile No: 8105895179
LinkedIn: https://in.linkedin.com/in/pramod.kn