Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
12
,
No.
3
,
Decem
ber
201
8
, p
p.
1340
~
1348
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
2
.i
3
.pp
1340
-
1348
1340
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Metric
s and
Ben
chma
rks for Emp
irical
and Comp
re
h
ension
Foc
us
ed Visu
alization
Re
search i
n the Sal
es Dom
ain
Loo Yew
Jie,
Do
ri
s
H
oo
i
-
Te
n W
ong,
Z
ari
na
M
at Z
ain, Nil
am Nur
Ami
r Sjarif,
R
osl
ina I
br
ah
im
,
Nu
r
az
ean Maaro
p
Univer
siti
Te
kno
logi
Ma
lay
sia
,
5
4100
Kuala L
um
pur,
Mal
a
y
s
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
ug
1,
2018
Re
vised
Oct
1
6
, 2
018
Accepte
d
Oct
30
, 201
8
Data
visualizati
on
is
an
eff
ort
which
ai
m
s
to
comm
unic
at
e
data
eff
ectiv
e
l
y
and
cl
e
arly
to
the
audi
en
ce
through
gra
phic
a
l
rep
rese
nt
at
ion
.
Dat
a
visual
izat
ion
eff
orts
m
ust
be
c
oordina
t
ed
with
an
und
ersta
ndi
ng
int
o
t
h
e
Cognit
ive
L
ea
r
ning
The
or
y
(
CLT
).
In
the
sale
s
dom
ai
n,
sale
s
data
visual
izat
ion
are
m
ade
poss
ibl
e
with
the
av
ai
l
able
Business
Inte
lligence
(BI
)
tool
s
such
as
Microsoft
Pow
er
BI,
T
abl
e
au,
Plo
tly
,
and
o
the
rs.
The
se
tool
s
al
low
se
amless
i
nte
ra
ct
ion
for
th
e
top
m
an
age
m
e
nt
as
wel
l
as
the
sale
s
for
ce
with
reg
ard
to
the
data.
Sale
s
dat
a
visualiz
at
io
n
comes
with
a
n
arr
a
y
o
f
adva
nt
age
s
such
as
self
-
servic
e
ana
l
y
sis
b
y
busi
ness
users,
rap
idly
ada
pt
t
o
cha
nging
busine
ss
condi
ti
ons,
a
nd
ena
b
le
conti
nuous
on
-
demand
rep
ort
ing
among
othe
rs.
The
adva
nt
age
s
of
sale
s
dat
a
visualization
al
so
comes
with
th
e
cha
l
l
enge
s
such
as
difficulty
in
i
dent
if
y
ing
visua
l
noise,
high
r
ate
of
imag
e
cha
nge
,
and
hig
h
per
form
ance
r
equi
rement
s.
In
an
eff
or
t
to
r
edu
ce
cogni
t
iv
e
ac
t
ivi
t
y
th
at
doe
s
not
enha
nce
learni
ng,
sal
es
visual
i
za
t
ion
dashboa
rd
m
ust
be
designe
d
in
a
wa
y
that
is
n
e
i
the
r
t
oo
sim
pli
stic
nor
too
complex
to
ensure
th
at
the
Intri
nsi
c
Cognit
ive
Loa
d
(I
CL),
Ext
r
insic
Cognit
ive
Loa
d
(ECL
),
an
d
Germ
ane
Cogni
ti
ve
Loa
d
(GCL
)
are
in
s
y
n
c
w
it
h
th
e
audienc
e
.
W
it
h
th
e
combinat
ion
of
sale
s
dat
a
v
isua
li
z
at
ion
and
C
L
T,
under
st
andi
n
g
complex
sale
s
detai
ls
qui
ckly
is
m
ade
poss
ibl
e
b
y
no
t
onl
y
the
top
m
ana
ge
m
ent
of
th
e
orga
nizati
on
,
bu
t
al
so
the sal
es
fo
rce
of
th
e
org
aniza
t
ion.
Ke
yw
or
d
s
:
Be
nch
m
ark
Cognit
ive L
oa
d
T
he
or
y
Data Vis
ualiz
at
ion
Me
tric
s
Sale
s
Copyright
©
201
8
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Doris
Hooi
-
Te
n Won
g
,
Un
i
ver
sit
i Te
knol
og
i M
al
ay
sia
,
54100 K
uala L
um
pu
r,
Ma
la
ysi
a
.
Em
a
il
:
do
risw
ong@
utm
.
m
y
1.
INTROD
U
CTION
The
pur
pose
of
data
visu
al
iz
at
ion
is
to
pro
je
ct
the
data
cl
early
an
d
e
ff
ec
tuall
y
to
the
s
pectat
or
s
by
us
in
g
grap
hica
l
illustrati
on.
It
is
a
cru
ci
al
par
t
of
the
pr
oces
s
to
un
c
overin
g
the
key
point
s
within
the
pr
ocess.
W
it
h
m
ulti
ple
source
of
da
ta
avail
able,
vi
su
al
iz
at
ion
is
i
m
po
rta
nt
an
d
is
bein
g
fu
l
ly
util
iz
ed
by
m
any
orga
nizat
ion
w
or
l
dw
i
de
in
m
akin
g
day
to
da
y
decisi
on
unti
l
i
t
is
reg
arded
as
an
vital
process
in
B
usi
nes
s
In
te
ll
igence
.
I
nfl
ux
of
data
oc
cur
s
com
m
on
ly
in
toda
y’s
da
ta
dr
ive
n
ec
os
y
stem
and
the
c
halle
ng
e
is
to
pres
ent
a
m
et
ric
and
ben
c
hm
ark
s
f
or
em
pirical
and
com
pr
e
hen
si
on
fo
c
us
e
d
visu
al
iz
at
ion
.
In
reali
ty
,
three
oth
e
r
i
m
po
rtant t
op
ic
s as s
uggeste
d by Sin
gh a
nd
Wajgi
[1
]
th
at
the
deci
sio
n
m
a
ker
s
w
il
l faces
su
c
h
as:
1)
The pr
oce
dure
of v
is
ualiz
at
ion
ca
n be
flexib
le
an
d
ve
rsati
le
.
2)
Suppor
ti
ng ev
i
den
ce
s ar
e
tra
nspare
nt to
acq
ui
re; an
d
3)
The
s
pee
d of c
om
pu
ti
ng
a
nd t
he
c
os
t
of
proc
essing.
This
pap
e
r
pr
e
sents
a
resea
rc
h
on
ho
w
wh
at
are
t
he
a
pprop
riat
e
m
e
tric
an
d
ben
c
hm
ark
s
in
pr
oducin
g
eff
ect
ive
v
is
ualiz
at
ion
in
the s
al
es dom
ai
n.
Trad
it
io
nally
, v
isuali
zat
ion h
as b
ee
n
the
do
m
ai
n
of
stat
ist
ic
s.
A st
an
dard
te
xtb
oo
k
in
sta
ti
sti
cs [
2] h
a
s
a
chap
te
r
on
cr
eat
ing
ba
r
cha
r
ts,
pie
char
ts,
l
ine
char
ts
,
hist
ogram
s,
et
c.
These
are
sim
ple
rep
r
esentat
io
ns
of
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Met
ric
s and Be
nchm
ar
ks f
or
Em
pirical a
nd
Compre
hensi
on
F
o
c
u
s
e
d
…
(
Loo Y
ew
Jie
)
1341
data
that
requi
re
sta
ndar
d
input.
H
ow
e
ve
r,
with
the
prolif
erati
on
of
ty
pe
s
and
var
ie
ty
of
data,
the
re
is
a
need
for
m
or
e
ty
pes
of
a
naly
ses
an
d
prese
ntati
on
s
that
a)
bri
ng
out
the
relat
io
nships
betwee
n
di
ff
ere
nt
el
em
en
ts
b)
su
m
m
arize
com
plex
data
with
sim
ple
and
e
asi
ly
un
de
rstoo
d
vis
uals
c)
si
m
pl
ify
the
vis
ualiz
at
ion
with
ou
t
the
loss
of
the
m
a
ny
dim
ension
s
of
the
data
an
d
d)
at
the
sam
e
t
i
m
e,
achie
ve
al
l
this
qu
ic
kly
with
easy
to
us
e
analy
ti
cal
too
ls.
Vis
ualiz
at
ion
is
par
ti
cula
rly
i
m
po
rtant
for
hi
erarch
ic
al
data,
w
her
e
the
in
div
id
ual
data
poi
nt
s
are
co
nnect
ed
i
n
a
tree
-
li
ke
st
ru
ct
ur
e,
with
l
arg
e
cl
us
te
rs
of
data
bro
ken
i
nto
s
ub
-
cat
e
gories.
T
he
hie
ra
rch
ic
al
analy
ses
of
da
ta
su
ggest
ed
he
re
can
he
lp
pe
op
le
to
see
r
el
at
ion
sh
i
ps
be
tween
var
ia
ble
s
and
gro
up
s
,
wh
il
e
m
aking
it
easy
to
chec
k
on d
at
a
ve
racit
y.
The
visu
al
iz
at
ion
hel
ps
to
un
derst
and
t
he
brea
k
-
up
o
f
sal
es
dat
a
int
o
cat
egories, s
ub
cat
egories et
c.
Be
nch
m
ark
in
g
ena
bles
c
om
pan
ie
s
to
see
t
heir
posit
ion
s
relat
ive
to
the
ir
com
petit
or
s
in
orde
r
to
exp
l
or
e
t
he
op
portu
niti
es
to
im
pr
ov
e
their
m
ark
et
posit
io
n.
T
his
is
ta
ke
n
by
it
s
de
finiti
on
s:
“B
enc
hma
rk
i
ng
i
s
the
process o
f
con
ti
nu
ously
m
easur
in
g
an
d
com
par
i
ng
o
ne
’s
busines
s
pro
cesses
against com
par
able
pr
ocesses
in
le
adin
g
org
anizat
ion
s
to
ob
ta
in
inf
orm
at
ion
that
will
help
t
he
org
anizat
ion
i
dent
ify
and
im
ple
m
ent
i
m
pr
ovem
ents”
[3
]
.
Wh
il
e
m
et
rics
can
be
def
i
ned
as
“Sta
nd
a
r
ds
if
m
easur
em
ent
by
w
hich
e
ff
i
ci
ency,
perform
ance,
pro
gr
e
ss, or
qu
a
li
ty
o
f
a
plan, p
ro
ces
s
or
pro
duct
can
b
e
asse
ssed [4
]
.
2.
RESEA
R
CH MET
HO
D
Me
lon
c
on
an
d
W
a
rn
e
r
[5
]
r
eviewe
d
the
m
ajo
r
cat
eg
or
i
es
f
ound
in
da
ta
visu
al
iz
at
ion
inclu
des
com
par
ison o
f t
ypes of
visu
al
i
zat
ion
s,
grap
hs, ic
ons,
o
t
her
a
nd onli
ne
.
2.1.
Co
m
pa
ri
s
on
of Ty
pes
o
f
V
is
ua
li
z
at
ion
1)
An
im
at
ion
s
an
d
sta
ti
c
visu
al
iz
at
ion
s
-
A
nim
at
ion
s
did
not
great
ly
pr
om
ote
po
sit
ive
le
arn
in
g
ou
tc
om
es,
and eve
n res
ulted in
p
e
rfo
rm
a
nce
degra
datio
ns
.
2)
Text,
ta
bles,
a
nd
ba
r
grap
hs
-
Gr
a
phs
a
re
great
ways
to
e
xpress
ris
k
c
om
m
un
ic
at
ion
pract
ic
e
due
t
o
t
heir
abili
ty
to
capt
ur
e
at
te
ntio
n
and
el
ic
it
inf
orm
ation
e
xtrac
ti
on
with
m
inim
al
cogniti
ve
ef
fort,
a
nd
w
il
l
i
m
pr
ove c
om
pr
ehensi
on.
3)
Tables
a
d
nd
bar
gra
ph
-
Wh
e
n
data
is
pr
ese
nted
in
these
fo
rm
ats,
aud
ie
nce
w
it
h
ex
pe
rience
an
d
knowle
dge
with
bar
gra
ph
s
prefe
rr
e
d
bar
graphs,
w
hile
th
os
e
with
e
xper
ie
nce
a
nd
ta
bl
es
f
ound
gr
a
phs
equ
al
ly
easy
to
us
e.
Wh
e
n
e
xam
ining
te
sts
with
bo
rd
e
rlin
e
resu
lt
,
bar
graphs
is
sti
ll
the
pr
e
ferred
m
edium
of v
is
ualiz
at
ion
.
4)
Nu
m
ber
s a
nd
i
cons
-
Gr
a
ph
ic
s an
d
ic
ons w
e
re th
e only
d
is
crep
a
ncy b
et
w
een im
pacted co
m
pr
ehe
ns
io
n and
recall
; bu
t
not i
m
pacted b
y t
he
actual
level
of icon
ic
it
y o
f gra
ph
ic
.
2.2.
Graphs
Gen
e
rall
y,
gra
ph
s
are
excell
ent
w
he
n
it
co
m
es
to
data
visu
al
iz
at
ion
,
al
thou
gh
ther
e
e
xist
a
de
ba
te
betwee
n
us
in
g
gr
ap
hs
an
d
li
nes.
H
oweve
r,
it
is
su
bj
ect
e
d
to
the
aud
ie
nces
li
te
racy
back
gr
ound.
It
is
al
so
disco
ver
e
d
that
gr
ap
h
co
nvent
ion
s
(tit
le
s,
le
ge
nds,
ori
entat
ion
a
nd
c
olo
r
s)
and
li
te
racy
rates
are
i
m
po
rtant
and
sh
oul
d be take
n
int
o
acc
ount.
2.3.
Ico
ns
Icons
a
re
a
n
e
ff
ect
ive
m
et
hod
to
s
how
i
nfor
m
at
ion
since
they
bo
os
te
d
recall
of
in
for
m
at
ion
an
d
eff
ect
ive
in
im
pro
ving
unders
ta
nd
in
g.
2.4.
Oth
ers
Othe
r
ty
pes
of
visu
al
iz
at
ion
include
s
pie
char
t,
m
aps
an
d
phot
ogra
ph
s
.
In
bri
ef,
pie
char
ts
we
re
pr
e
ferred
w
he
n
dis
play
ing
ge
no
m
ic
risk
i
nfor
m
at
ion
due
t
o
the
sim
i
la
rity
to
c
omm
on
ob
j
ect
a
nd
the
se
e
m
ing
si
m
plici
t
y
of
ba
sic
per
ce
ntag
es
and
al
lo
we
d
sim
pler
visua
li
zat
ion
.
Stu
di
es
sh
owe
d
tha
t
do
m
ai
n
kn
owle
dge
can
in
flue
nce
i
nfor
m
at
ion
sel
ect
ion
a
nd
un
de
rstan
ding
of
c
om
p
le
x
grap
hi
cs,
an
d
t
hey
of
fer
em
pirical
su
pp
or
t
for
the
d
at
a
vis
ualiz
at
ion
c
onc
ept that t
he display
sho
uld
a
void i
nclu
de
a
ny
m
or
e in
form
at
ion
that
is re
qu
i
red.
2.5.
Onli
ne
The
th
ree
m
os
t
widely
discu
ssed
ones
i
nclud
e
s
pe
rsonal
healt
h
rec
ord,
patie
nt
inf
or
m
at
ion
we
bs
it
e
and
el
ect
r
onic
healt
h
rec
ord
. W
hile
they
are
su
bject
to
t
heir
res
pecti
ve
int
erf
ace desig
ns
,
the
m
ajo
r
co
nc
ern o
f
these
vis
ualiz
at
ion
s
is
t
hat
the
gr
a
phic
al
inf
orm
ation
wer
e
t
oo
com
plex
a
nd
incl
ud
e
d
e
xc
essive
in
f
or
m
a
ti
on
t
o
abs
orb
a
nd un
de
rstan
d.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1340
–
1348
1342
2.6.
Visua
l
Analy
ti
cs (VA
)
Figure
1
Vis
ua
l
An
al
yt
ic
s
(VA)
c
ould
be
a
t
each
a
pp
ea
rin
g
crit
ic
al
gu
ara
nt
ee
in
m
aking
a
diff
e
re
nce
cl
ie
nts
pick
up
kn
ow
le
dg
e
i
nto
c
om
plex
inf
or
m
at
ion
.
V
A
de
vices
use
hu
m
an
per
c
e
ptu
al
an
d
s
ubje
ct
ive
capaci
ti
es
by
ut
il
iz
ing
intel
li
gen
t
re
pr
ese
ntati
on
s
as
i
nterf
a
ces
am
on
gs
t
cl
ie
nts
an
d
their
inf
or
m
at
ion
,
in
thi
s
way m
aking
in
form
ation
relat
ed unde
rtakin
gs
m
or
e com
pelli
ng
a
nd ef
fecti
ve.
Figure
1. O
verview
of
visu
al
analy
ti
cs
Visu
al
A
naly
ti
c
exp
ect
s
to
de
crease
com
plex
intel
le
ct
ual
work
t
o
proce
s
s
huge
am
ou
nt
of
data
set
s
towa
rd
s
an an
s
wer
a
ble in
form
at
ion
[6
]
.
The
inf
or
m
at
i
on
f
r
om
co
m
p
any’s
operati
on
with
cust
ome
rs’
interact
io
n
are
ve
ry
rich.
The
re
ar
e
so
m
e
structur
e
d
data
w
her
e
can
be
sto
re
d,
retrieve
an
d
analy
ze
in
sp
r
eads
heets
or
i
n
relat
ion
al
da
ta
base.
Ther
e
are
al
s
o
sem
i
-
structu
re
d
data
li
ke
em
ai
l
data
or
we
bsi
te
traff
ic
date
w
her
e
nee
d
e
xtra
e
ffor
t
t
o
proces
s
and
a
naly
ze
then
su
m
m
arize
it
in
sign
ific
an
t
ways.
For
un
structu
re
d
data
wh
e
re
it
is
known
as
a
ve
ry
wealt
h
of
data
wh
ic
h
are
relat
ed
t
o
c
om
pan
y;
custo
m
ers,
re
views
,
te
stim
on
ia
ls,
and
so
ci
al
m
edia.
It
is
im
po
rtant
f
or
the
com
pan
y
to
ha
ndle
the
da
ta
,
storing,
ret
rievin
g
a
nd
m
anag
i
ng
al
l
different
ty
pe
of
da
ta
becau
se
it
is
help
the co
m
pan
y t
o pr
i
or
it
iz
e the
perform
anc
e m
easur
e
s
based
on these
d
at
a.
To
pe
rfo
rm
th
e
ben
c
hm
ark
and
m
easur
em
e
nts
f
or
in
form
at
ion
re
pr
e
sent
at
ion
,
orga
nizat
ion
util
iz
e
s
excel s
pr
ea
dsh
eet
an
d t
ablea
u t
oo
l i
nco
m
e in
form
ation
:
1)
Dev
el
op
a
pe
r
iod
a
rr
a
ngem
e
nt
plo
t
of
nu
m
ber
of
re
que
sts
pu
t
for
c
onsist
ently
in
t
he
in
form
at
ion
al
ind
e
x.
2)
Visu
al
iz
e the
total
num
ber
of
requests
put f
or e
ver
y
day of t
he
m
on
th.
3)
Show
a
gu
i
de
represe
ntati
on
with
e
ver
y
on
e
of
the
sta
te
s
in
the
U
S
a
nd
qu
al
it
ie
s
f
or
the
qu
a
ntit
y
of
requests
put i
n ea
ch
sta
te
a
nd the
a
ver
a
ge
i
nc
om
e p
er arra
ng
e in that
sta
te
.
4)
Gr
a
ph the
qua
ntit
y of
sit
e
vis
it
s ev
ery
day f
or all
d
at
es i
n
t
he data
set
.
5)
Gr
a
ph the
qua
ntit
y of
sit
e
hits f
or
al
l dates i
n
the
d
at
aset
.
6)
Desig
n
a
das
hboa
r
d
that
al
l
t
he
wh
il
e
s
how
s
the
guide
pe
r
cepti
on
a
nd
th
e
diag
ram
with
the
quantit
y
of
s
it
e h
it
s for
all
dates in
the
dat
aset
.
7)
Desig
n
Strate
gy
m
ap
and
bal
anced
sc
or
eca
r
d.
A
strat
egy
m
ap
is
a
sup
portive
re
pr
e
sen
ta
ti
on
a
pp
a
ratu
s
worked
ar
ound
the
bala
nced
scor
eca
rd
id
e
as
that
ou
tl
ine
s
ci
rcu
m
sta
nces
and
e
nd
res
ul
ts
con
necti
ons
betwee
n
key
act
ivi
ti
es
disp
la
ye
d
cl
os
e
by
ben
chm
ark
and
m
easur
em
ents
m
ark
er
s.
Re
gu
la
rly
,
a
te
chn
iq
ue o
utli
ne fo
ur p
a
rtic
ul
ar terr
it
ori
es
f
or m
easur
em
en
ts an
d ben
c
hma
rk assessm
ent.
8)
Finan
ci
al
po
i
nt
of
view
–
dem
on
strat
es
a
ppr
oach
es
t
o
accom
plish
econom
ic
dev
el
opm
ent
to
fu
lfi
l
l
inv
est
or
s
(
sla
c
k
m
ark
ers
).
9)
Custom
er
po
in
t
of
view
–
portrays
acc
ompli
sh
m
ent
wit
h
cl
ie
nts
and
char
act
e
rizes
cl
ie
nt
sect
ion
s
(a
blen
d of sla
ck a
nd lead m
ark
e
rs)
.
10)
In
te
r
nal
proces
s point
of v
ie
w
–
e
xhibit
s
how
estee
m
is con
ve
ye
d
to cli
e
nts
(lead m
ark
ers
).
11)
Le
arn
i
ng
a
nd
dev
el
op
m
ent
point
of
vie
w
–
center
s
ar
ound
i
nd
i
viduals,
innov
at
io
n,
a
nd
hiera
rch
ic
a
l
atm
os
ph
e
re
(lead m
ark
ers
).
12)
Pr
ope
rtie
s
of
m
et
rics.
To
outl
ine
KP
I
s,
it
is
us
e
fu
l
t
o
re
m
e
m
ber
that
al
l
tog
et
he
r
f
or
a
m
et
ric
to
be
fruit
fu
l,
it
ought
to
be
Si
m
ple
to
co
m
pr
ehe
nd
a
nd
be
nc
hm
ark
against;
Ma
p
to
key
business
exe
rcises,
act
ivit
ie
s
to
c
om
es
abo
ut;
Acti
on
a
ble
–
center
co
ns
id
e
rati
on
a
nd
gu
i
de
rig
ht
co
nduct;
Re
li
able
and
su
bst
antia
l;
and Tim
el
y (SMART)
.
13)
Dashb
oards
-
Orgaiza
it
ion
r
egu
la
rly
util
ize
el
ect
ro
nic
da
shbo
a
rds
to
see
KP
Is
.
A
da
shbo
a
rd
viabl
y
portrays
m
ark
ers
util
iz
ing
des
ign
s
wh
ic
h
m
a
kes
it
consi
de
r
ably
le
ss
dem
a
nd
i
ng
t
o
rec
ou
nt
a
story
a
nd
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Met
ric
s and Be
nchm
ar
ks f
or
Em
pirical a
nd
Compre
hensi
on
F
o
c
u
s
e
d
…
(
Loo Y
ew
Jie
)
1343
convey
it
al
l
throu
gh
t
he
orga
nizat
ion
. I
t
can
li
kew
ise
be
furn
is
hed
w
it
h
noti
ce
sign
s or
a
la
rm
s
con
veye
d
wh
e
n
a
m
et
ric
is o
utside
of pr
eset
p
aram
et
ers.
Sale
s
forecast
ing
is
am
on
g
t
he
f
undam
ental
inp
uts
f
or
pl
ann
i
ng
decisi
ons
th
rou
ghout
the
supp
ly
chain.
Esti
m
at
i
ng
fu
t
ur
e
dem
and
m
or
e
accu
ratel
y
is
critica
l
fo
r
m
eet
ing
i
t,
wh
il
e
m
ini
m
iz
ing
inve
nt
ory
and
oth
e
r
relat
e
d
c
os
ts.
These
d
e
m
and
esti
m
at
e
s ar
e
oft
en
m
od
el
le
d base
d o
n hist
or
ic
al
pat
te
rn
s i
n
the
dat
a [
7].
3.
FIN
DINGS
Accor
ding
to
Hav
em
o
[
8],
f
ro
m
a
rep
ort
in
g
pe
rs
pecti
ve,
visu
al
m
eans
su
c
h
as
gra
phs
a
nd
oth
e
r
visu
al
isa
ti
on
is
essenti
al
in
increasin
g
bu
s
iness
m
od
el
s
pr
ese
ntati
on.
Accor
ding
to
the
2015
Glea
ns
ig
ht
Be
nch
m
ark
Re
port
[
9]
on
data
visu
al
iz
at
ion,
there
ar
e
a
num
ber
of
reas
on
s
w
hy
data
vis
ualiz
at
ion
will
be
i
m
ple
m
ented.
3.1.
Empowe
r n
on
-
IT
P
r
of
es
sion
als
If
a
vaila
ble
to
ols
are
to
o
c
om
plex,
it
’s
ve
r
y
com
m
on
for
or
ga
nizat
ion
s
t
o
dep
e
nd
heavi
ly
on
I
T
f
or
run
ning
querie
s,
custom
iz
ing
repor
ts,
a
nd
c
onduct
ing
a
nal
ysi
s.
All
these
things
create
bott
le
neck
s
for
us
ers
.
More
an
d
m
or
e
com
pan
ie
s
are
lookin
g
to
increase
ad
opti
on
of
sel
f
-
s
erv
ic
e
BI
to
s
upport
their
goal
in
em
pow
e
rin
g n
on
-
I
T
prof
es
sion
al
s
.
3.2.
Rapidl
y
Adap
t
t
o Ch
angin
g B
usiness
Con
ditions
Data
vis
ualiz
at
ion
is
i
deal
f
or
arti
culat
in
g
qual
it
at
ive
ch
ang
e
s
in
busi
ness
data
set
s
su
c
h
a
s
a
n
acqu
isi
ti
on,
m
erg
e
r,
ne
w
bus
iness
unit
,
or
change
in
the
data
hiera
rch
y.
Data
visu
al
iz
at
ion
m
ay
pr
ovide
a
gr
eat
way to
und
e
rstan
d va
riances i
n
the
nu
m
ber
s w
it
h gr
e
at
er ease.
3.3.
Encour
age
Data Expl
oratio
n
Giving
us
e
rs
visu
al
ly
stim
ulati
ng
,
a
nd
sim
ple
inter
faces
m
ini
m
iz
es
the
sk
il
ls
re
quire
d
to
co
nduc
t
analy
sis.
To
p
Perfo
rm
ers
reco
gniz
e
that
the
best
thin
g
the
y
can
do
f
or
th
e
bu
si
ness
is
gi
ve
us
e
rs
with
c
on
te
xt
about
how
t
o
i
nter
pr
et
data tr
en
ds ea
sy ac
ce
ss to
t
he data
[9].
Ali et
.al
.
[
10]
e
m
ph
asi
zed som
e o
f
t
he big
da
ta
v
isuali
zat
io
n pro
blem
, w
hich
incl
ud
e:
1)
Visu
al
noise
:
High
relat
ivit
y
between
eac
h
obj
ect
s
in
the
dataset
,
resu
lt
ing
hi
gh
di
ff
ic
ulty
to
separ
at
e
them
.
2)
Inform
at
ion
lo
ss:
So
m
e
inf
orm
at
ion
are
sac
rificed
i
n
t
he
e
ffor
t
t
o
im
pr
ov
e
dataset
visibi
li
ty
and
inc
rea
se
respo
ns
e ti
m
e.
3)
Pers
on
al
pe
rce
ption an
d
i
nter
pr
et
at
io
n of
t
he
v
is
ual
isa
ti
on
.
4)
Highly
dynam
ic
data
requires
co
ns
ta
nt
vis
ua
li
sat
ion
up
dating
inc
reases
di
ff
ic
ulty
f
or
us
e
r
to
react
t
o
th
e
fig
ur
es
sho
wn.
5)
High
perform
a
nce
requirem
ents:
Dy
nam
ic
visu
al
iz
at
ion
dem
and
s
for
m
or
e
re
qu
i
rem
e
nts
c
om
par
ed
to
sta
ti
c v
isuali
zat
ion
.
Ma
ny
too
ls
ha
ve
bee
n
in
ve
nted
to
help
us
out
from
the
abo
ve
pro
blem
.
The
m
os
t
cru
ci
al
featur
e
tha
t
a
vis
ualiz
at
ion
m
us
t
ha
ve
is
interact
ivit
y.
I
n
t
he
business
w
or
l
d,
m
any
orga
nizat
ion
ha
ve
opte
d
f
or
visu
al
iz
at
ion
t
oo
ls
to
m
ake
interest
ing
das
hboa
rd
a
nd
at
tract
ive
pr
e
sen
ta
ti
on
s.
Am
ong
the
m
os
t
popu
la
r
visu
al
iz
at
ion
t
oo
ls
a
re
s
ummari
zed
in
Tabl
e
1.
Ali
et
.
al
.
[10]
com
par
ed
these
to
ols
on
the
basis
of
va
rio
us
at
tribu
te
s. Som
e of the
consid
erati
on
s
whe
n choosi
ng the
righ
t
vis
ualiz
at
ion
to
ols a
re list
ed belo
w:
1)
To
ol is
ope
n
s
ource
or
no
t.
2)
Visu
al
isa
ti
on c
reated all
ows
use
r
to
interact
with them
.
3)
Su
it
able cl
ie
nt
ty
pe
or
pack
a
ge
s to
c
reate t
he
v
is
ualisa
ti
on
.
4)
Re
adiness
to
i
nt
egr
at
e
with
da
ta
sour
ces
su
c
h as H
ad
oop Hi
ve,
G
oogle
Anal
yt
ic
s,
et
c.
5)
Av
ai
la
bili
ty
o
f
tutor
ia
ls t
hr
oug
h
Ma
ssiv
e
Op
e
n On
li
ne
Co
urs
es (
MO
OCs
).
6)
Accessibil
it
y and avail
abili
ty
o
f
App
li
cat
i
on
Pr
og
ram
m
ing
Interface
(AP
I).
Table
1.
C
om
par
iso
n of so
ftw
are att
rib
utes
use
d
i
n data vis
ualiz
at
ion
sales
dom
ai
n
Tableau
Po
wer
BI
Plo
tly
Gep
h
i
Excel 2
0
1
6
Op
en
Sou
rce
N
N
Y
Y
N
Interactive
Y
Y
N
N
Y
Desk
to
p
Clien
t
Y
Y
N
Y
Y
On
lin
e Clien
t
Y
Y
Y
N
Y
Mob
ile App
.
Y
Y
N
N
Y
Integ
ration
Y
Y
N
N
Y
MOOCs
Y
Y
Y
Y
Y
API
Y
Y
Y
Y
Y
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1340
–
1348
1344
Althou
gh
t
he
a
forem
entioned
too
ls
offe
rs
powe
rful
feat
ures
and
oft
en
use
d
by
bu
si
nes
ses,
ho
wev
e
r
they
also c
om
e
w
it
h
li
m
it
s
/dem
erit
s as h
ig
hl
igh
te
d by t
he
a
uthors.
1)
Tableau:
Ta
ble
au
P
ub
li
c
only
com
es
with
a
1G
B
st
or
a
ge
a
nd
for
la
r
ger
w
ork
re
quirem
e
nts,
li
cense
of
the
serv
e
r
a
nd Ta
bl
eau D
e
sk
t
op
will
b
e
require
d.
2)
Mi
cro
s
of
t
P
ow
er
BI:
It
c
om
es
with
a
fr
ee
ve
rsion
bu
t
us
e
rs
m
us
t
hav
e
a
Wor
k
acc
ount
and
it
is
li
m
it
e
d
t
o
250
MB
of sto
r
age
for wo
r
kbook.
It is also
slow
e
r
if
co
m
pared to
Ta
bleau.
3)
Plotl
y:
Pr
o
us
e
rs
ha
ve
lim
it
ed
to
on
ly
50
0
K
B
fo
r
uploa
d
s
iz
e.
Even
i
f
pr
of
essi
onal
ver
s
ion
,
y
ou
will
get
un
li
m
i
te
d
chart
s
bu
t
uploa
d
s
iz
e
of
file
s
will
be
lim
it
ed
to
on
ly
5
MB
.
P
rogr
am
m
ing
sk
il
ls
are
requir
e
d
and no
offici
al
offli
ne
cl
ie
nt
f
or Plotly
is ava
il
able.
4)
Gephi:
Only
s
pe
ci
al
iz
es in gra
ph v
is
ualiz
at
ion
, ca
nnot
be
a
ppli
ed fo
r othe
r t
ypes of
visu
al
i
zat
ion
s.
5)
Excel
2016:
Mi
cro
s
of
t
Office
is
a
pai
d
a
pp
li
cat
io
n
an
d
the
on
ly
the
Office
36
5
s
ubscri
ber
s
will
gai
n
acce
ss to
t
he A
PI
.
Re
fer
ri
ng
to
Ma
gee
et
.al.
[
11
]
pro
per
dat
a
visu
al
iz
at
io
n
increase
s
the
abili
ty
of
the
s
al
esper
s
on
to
interp
ret
the
da
ta
visu
al
iz
at
ion
p
rese
nted.
The
pa
per
al
s
o
sta
te
s
that
t
he
hu
m
an
bra
in
is
ha
r
d
-
wir
ed
to
narrati
ve
an
d
vi
su
al
patte
rn
s
a
nd
no
t
m
at
he
m
at
ic
al
on
es.
Prop
e
r
ide
ntific
at
ion
of
sal
espe
r
so
n
inte
rest
or
fo
c
us
is
al
so
im
po
rta
nt
in
prese
ntin
g
t
he
sal
es
data
to
t
he
sal
es
person.
Be
si
des
t
ha
t,
data v
is
ualiz
at
ion
i
n
this
c
on
te
xt
is
al
so
a
n
orga
nizat
ion
al
c
ha
nge
a
ge
nt.
Sale
s
per
s
on
wer
e
a
ble
to
ide
ntify
their
key
f
oc
us
in
orde
r
to
ide
ntify
the r
i
gh
t l
ea
ds
to brin
g
i
n
the
sal
es as
discusse
d
i
n
the
sam
e
p
a
per.
4.
PROP
OSE
D SOLUTI
ON
It
is
un
de
nia
ble
that
or
ga
niz
at
ion
s
norm
al
l
y
hav
e
a
colle
ct
ion
of
datab
ase,
w
her
e
ea
ch
data
bas
e
storing
diff
e
re
nt
piece
of
in
form
ation
.
Howev
e
r,
visu
al
i
zi
ng
the
se
hu
ge
ch
uc
k
of
i
nfor
m
at
ion
is
us
ua
ll
y
chall
eng
i
ng
an
d
m
igh
t
le
ad
to
co
nfusi
on
i
f
no
t
pr
ese
nte
d
appr
opriat
el
y.
Gen
e
rall
y,
nu
m
ber
s
and
fig
ur
es
by
them
sel
ves
do
no
t
car
ry
m
uc
h
m
eaning
unle
ss
represe
nted
us
in
g
the
rig
ht
visu
al
.
The
quest
io
n
he
re
is,
whe
n
it
co
m
es
to
vi
su
al
iz
at
ion
,
es
pecial
ly
in
the
sal
es
do
m
ai
n,
wh
at
are
the
m
et
ric
and
be
nch
m
ark
one
sh
oul
d
fo
ll
ow
t
o
m
ak
e
rep
ort
in
g
w
ork
eff
ect
i
ve
and
easi
ly
under
st
ood
w
he
n
pr
ese
nted
to
the
m
anag
em
e
nt
or
sta
keholde
rs
?
This
sect
ion
discusse
s
so
m
e
of
the
propose
d
m
e
tric
a
nd
ben
c
hm
ark
fo
r
em
pirical
and
com
pr
ehe
ns
io
n f
ocu
se
d vis
ualiz
at
ion
in
the s
al
es dom
ai
n.
This
pa
per
a
dopts
the
fou
ndat
ion
for
the
desig
n
of
ins
tructi
on
an
d
a
ssessm
ent
as
pro
po
se
d
by
Lep
pink
[12]
,
wh
ic
h
ai
m
s
t
o
kee
p
co
gn
it
ive
act
ivit
y
to
it
s
m
ini
m
al
si
nce
it
will
j
eop
ar
dize
le
arn
in
g.
T
hi
s
fr
am
ewo
r
k
re
volves
ar
ound
the
Co
gn
it
ive
Loa
d
The
or
y
as
the
dev
el
opm
e
nt
and
aut
om
at
ion
of
co
gnit
ive
schem
as
reg
a
r
ding
c
onte
nt
t
o
be
delive
re
d
a
nd
le
ar
nt
by
th
e
au
dience
.
T
he
three
ty
pes
of
c
ogniti
ve
l
oa
d
a
re:
In
tri
ns
ic
Co
gnit
ive Lo
a
d (I
CL
),
E
xtra
ne
ou
s
Cognit
ive L
oa
d (ECL)
and
G
erm
ane Co
gnit
ive L
oad (
GCL
).
Wh
e
n
pr
e
par
i
ng
a
p
resen
ta
ti
on
dec
k
to
re
port
num
ber
and
fig
ur
es
,
it
i
s
i
m
po
rtant
to
ensure
it
is
desig
ne
d
in
suc
h
a
way
that
on
ly
a
m
ini
m
um
of
work
i
ng
m
e
m
or
y
po
we
r
is
required
for
co
gn
it
ive
pro
cesses
that
do
not
co
ntribute
to
le
arn
in
g
as
m
uch
.
Ba
la
nce
is
the
key
in
this
sit
uation
w
her
e
th
e
pr
ese
ntati
on
dec
k
sh
oul
d
co
ns
ist
of
el
em
ents
that
are
cl
ear
an
d
easi
ly
under
s
tood.
Mo
re
ov
e
r,
in
re
portin
g
nu
m
ber
s
a
nd
figures
,
it
is
no
t
a
go
od
pr
act
ic
e
to
m
erely
le
arn
the
ste
ps
of
a
proce
dure.
Ra
ther
,
they
hav
e
to
be
under
ta
ken
in
a
par
t
ic
ula
r
se
quence
to
e
nsu
re
a
c
orrect
s
olu
ti
on
for
a
giv
e
n
sit
uatio
n.
The
seq
ue
nce
m
at
te
rs
and
that
interact
ivit
y
add
s
to
ICL.
Ta
ke
the
case
of
a
bu
sin
ess
anal
yst
,
in
su
ch
a
sit
uation,
hav
i
ng
to
a
ddress
a
root
-
cause
analy
sis
on
a
drop
in
sal
es,
wh
ere
there
a
re
m
any
po
ssible
dia
gnos
es
m
ay
tak
e
the
ICL
for
le
ss
exp
e
rience
d
a
naly
st
to
the
lim
it
s
of
their
work
i
ng
m
em
or
y.
This
will
le
ads
to
cre
at
ing
a
vis
ualiz
at
ion
das
hboard
t
hat
is
ove
r
sim
pli
sti
c
and
fail
t
o
delive
r
up
to
the
be
nch
m
ark
.
O
n
t
he
c
on
t
ra
ry,
a
m
or
e
a
dv
anced
analy
st
will
ev
entuall
y
ex
per
i
ence
a
l
ow
e
r
I
CL
in
s
uch
a
s
it
uation
beca
use
they
can
act
ivate
m
or
e
de
ve
lop
e
d
and
pe
r
hap
s
al
read
y
m
or
e
a
ut
om
a
te
d
co
gnit
ive
sc
hem
as
t
han
their
le
ss
exp
e
rience
d
c
ollea
gu
e
s.
Thi
s
will
le
ads
to
creati
ng
a
vis
ualiz
at
ion
das
hboa
r
d
that
is
ov
er
co
m
plex
and
hard
to
be
unde
rs
tood
if
the
audi
ence
do
e
s
no
t
ha
ve
t
he
sam
e level
o
f
I
CL.
H
e
nce,
careful
ref
le
ct
ion o
n
t
his I
CL
factor i
s
of
pa
r
a
m
ou
nt im
po
rt
ance.
Herna
ndo
et
.
al
.
[
13
]
c
oncl
ud
e
d
t
hat
it
i
s
not
a
ppr
opr
ia
te
to
s
how
al
l
the
d
e
pe
ndencies
an
d
interrelat
io
ns
hi
ps
that
exist
in
big
data
dom
a
ins,
beca
us
e
th
ere
w
ou
l
d
be
a
n
excess
of
i
nfor
m
at
ion
that
woul
d
m
ake
it
i
m
po
ssible
to
detect
the
releva
nt
r
esults.
I
n
gen
e
ral,
an
orga
niz
at
ion
ca
n
co
nsi
der
the
f
ollo
wing
pip
el
ine
pro
po
sed by Si
n
gh a
nd
W
a
jgi [
1] a
s d
e
picte
d
i
n
Fi
gure
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Met
ric
s and Be
nchm
ar
ks f
or
Em
pirical a
nd
Compre
hensi
on
F
o
c
u
s
e
d
…
(
Loo Y
ew
Jie
)
1345
Figure
2. Sale
s
d
at
a
visu
al
iz
at
ion
pip
el
in
e
4.1.
D
ata
P
arser
Dep
e
ndin
g
on
sit
uations
,
us
er
m
ay
find
data
set
with
m
ultip
le
entries
to
be
releva
nt
or
irreleva
nt
.
Ther
e
f
or
e,
pa
rs
ing
will
be
perform
ed
in
j
ava
us
in
g
j
a
va.uti
l.Iterat
or
cl
ass
to
exam
ine
the
featur
e
s
that
exist
i
n
the d
at
a
set
.
4.2.
D
ata Cle
aner
It
is
necessary
to
be
rem
ov
ed
and
cl
eane
d
f
r
om
the
dataset
to
keep
them
relevan
t
to
the
sit
uation
an
d
reduce
unneces
sary com
pu
ta
ti
on r
es
ources
.
4.3.
D
ata
Tr
an
s
fe
r HSS
F
Wor
kbooks
w
e
re
c
hosen
f
or
s
toring
t
he
Fil
eIn
putSt
ream
prov
i
ded
by
the
us
er
f
or
c
hang
e
the
featu
re
nam
e
exist
in
t
he
data
set
.
Th
e
nam
es
of
the
featur
e
m
ay
need
furthe
r
effor
t
to
recti
fy
so
there
are
in
pro
per
form
at
m
a
y
not
be
i
n
proper
form
a
t.
Fo
r
i
nst
ance,
Purcha
s
e
I
d
will
be
e
xpress
ed
as
P
uI
D
wh
ic
h
m
ay
cause
conf
us
io
n.
4.4.
D
atabase
On
ce t
he data
are
pro
pe
rly
p
r
ocesse
d,
it
w
il
l be im
po
rted
i
nt
o
the
database
w
hic
h
c
on
ta
in
ap
pr
opriat
e
data rele
van
t t
o
the
user i
n
th
e pro
per f
or
m
at
.
4.5.
C
ache
Ca
c
he
is
of
te
n
us
ed
to
f
requ
ently
us
ed
dat
a
that
is
extracte
d
from
the
database
to
reduce
tim
e
and
effor
t t
o
re
peat
edly
p
e
rfor
m
ing
the
sam
e extr
act
ion
.
4.6.
Visu
aliz
at
ion
Ti
m
e
du
rati
on
pro
vid
e
d
by
t
he
e
nd
use
r
is
usual
ly
sp
eci
f
ie
d
w
he
n
it
co
m
es
to
data
vi
su
al
iz
at
ion.
High
val
ue
cu
s
tom
ers,
reg
io
na
l
sal
es
and
t
op
pr
oducts
can
be
vis
ualiz
ed.
By
us
ing
t
his
pract
ic
e,
en
d
us
er
will
then
car
ried
out
their
res
pe
ct
ive
decisi
on
m
aking
proc
ess
Howe
ver,
wh
e
n
desi
gning
the
vis
ual
iz
at
ion
das
hboard,
it
is
i
m
po
rtant
to
rem
e
m
ber
no
t
to
inc
orp
or
at
e
un
neces
sary
ECL
and
aud
ie
nce
s’
le
vel
of
knowle
dge s
ho
uld
be
ta
ken in
to acco
unt.
5.
APPLI
CA
TI
ON
In
this
sect
io
n
how
us
er
c
an
ap
ply
this
theor
y
in
sal
es
data
visu
al
iz
at
ion
is
discuss
e
d.
T
he
appr
opriat
e
gra
ph
s/c
har
ts
m
ust
be
a
pp
li
ed
in
the
su
it
a
ble
co
ntext
to
incr
ea
se
ICL
a
nd
ev
entuall
y
boost
GCL.
Ab
el
a
[14]
sum
m
arized
a
char
t
sugg
e
sti
on
s
that
is
com
pact
fo
r
vis
ualiz
at
ion
use
.
I
n
ge
ner
al
,
t
her
e
a
r
e
fou
r
cat
egories o
f
c
har
t,
which
inc
lud
es:
c
om
par
ison,
relat
ion
s
hi
p,
distrib
utio
n an c
om
po
sit
i
on.
Com
par
ison
graphs
can
be
f
ur
t
her
s
ubdiv
i
ded
i
nto
tw
o
s
m
al
le
r
gr
oup,
ei
ther
they
are
com
par
ing
a
m
on
g
it
em
s
or
o
ve
r
ti
m
e.
For
c
om
par
ison
a
m
on
g
it
em
s,
us
er
ca
n
c
onsider
va
riable width
colum
n
c
har
t,
table
with
em
bed
de
d
char
ts
,
bar
c
har
t
or
colum
n
char
t.
Fo
r
co
m
par
ison
s
ove
r
tim
e,
us
er
can
co
ns
ide
r
ci
rcu
la
r
area
char
t,
li
ne
c
hart
,
colum
n
chart
or
li
ne
char
t
.
Re
la
ti
on
sh
ip
betwee
n
tw
o
va
riables
can
be
expresse
d
in
s
cat
te
r
plo
t.
Me
a
nwhi
le
,
relat
ionshi
p
with
t
hr
ee
va
r
ia
bles
can
be
e
xpresse
d
i
n
bubb
le
c
ha
rt.
Dist
ribu
ti
on
with
sing
le
var
ia
ble
can
be
expresse
d
in
colum
n
histo
gra
m
fo
r
few
dat
a
po
i
nts
an
d
li
ne
hist
ogram
fo
r
m
any
data
points.
Distrib
ution
wi
th
tw
o
var
ia
bles
can
be
ex
pre
ssed
i
n
scat
te
r
plo
t.
Finall
y,
di
stribu
ti
on
with
th
ree
va
riabl
es
can
be
e
xpresse
d
i
n
3D
A
rea
C
ha
rt.
Com
po
sit
ion
w
hich
are
c
hangin
g
over
tim
e
can
be
ex
pr
ess
ed
us
i
ng
sta
cke
d
100%
c
olu
m
n
char
t,
sta
cke
d
c
olu
m
n
chart
,
sta
cke
d
10
0%
area
c
har
t
or
sta
cke
d
are
a
char
t
.
Wh
il
e
sta
ti
c
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1340
–
1348
1346
com
po
sit
ion
c
an
be
e
xpress
ed
us
i
ng
pie
char
t,
wate
rfal
l
char
t
or
s
t
acked
100%
colum
n
chart
with
su
bc
om
ponen
t
s.
The
a
ppr
opriat
e
cha
rt
m
us
t
be
a
pp
li
e
d
to
the
ri
gh
t
c
onte
xt
to
im
pr
ov
e
GCL
a
nd
a
vo
i
d
c
reati
ng
unnecessa
ry E
CL. Fig
ure
3
s
umm
arizes t
he c
har
t s
ugge
sti
on a
nd their
crit
eria.
In
a
nother
co
ntext,
visu
al
iz
ing
sal
es
rel
at
ed
ge
ogra
phic
or
dem
og
ra
phic
data
in
m
aps
does
no
t
necessa
ry
le
ad
s
to
bette
r
ICL
since
not
e
very
on
e
has
the
s
a
m
e
le
vel
of
ge
ogra
ph
ic
al
know
le
dg
e
.
In
Fi
gure
4,
the av
e
ra
ge bir
th r
at
e
for
c
ount
ries in the
r
e
gi
on of
Asia a
nd
The Am
ericans ar
e c
om
par
ed
us
in
g
m
aps.
Figure
3. Cha
rt Sug
gestio
n
Figure
4. Vis
ua
li
zi
ng
b
i
rth ra
te
b
y re
gion
usi
ng
m
aps
Althou
gh
aest
hetic
an
d
a
pp
eal
ing
,
ho
wever,
t
his
will
c
reate
ECL
if
the
au
die
nce
geog
raphical
knowle
dge
is
l
i
m
i
te
d.
It
is
bette
r
to
re
pr
ese
nt
the
c
om
par
iso
n
of
ave
ra
ge
birth
rate
betwee
n
tw
o
re
gions
us
i
ng
a
sim
ple
bar
grap
h
si
nce
it
is
cl
ear
a
nd
easi
ly
underst
ood.
I
n
Fi
gure
5,
one
ca
n
ea
sil
y
con
cl
ud
e
that
Asian
countries
ha
ve
a
bette
r
aver
age
bi
rth
r
at
e
than
The
Am
ericans
c
ountries.
Audi
ence
with
ou
t
m
uch
geog
r
ap
hical
knowle
dge
can
easi
ly
visu
al
iz
e
the
nu
m
ber
and
fig
ur
es
,
he
nce
le
adin
g
to
bette
r
ICL
an
d
GCL
and av
oid
i
ng E
CL.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Met
ric
s and Be
nchm
ar
ks f
or
Em
pirical a
nd
Compre
hensi
on
F
o
c
u
s
e
d
…
(
Loo Y
ew
Jie
)
1347
Figure
5. Vis
ua
li
zi
ng
b
i
rth ra
te
b
y re
gion
usi
ng
ba
r
c
har
ts
6.
CONCL
US
I
O
NS
A
ND FUT
U
RE
WO
RKS
As
the
ne
ed
for
quic
k
decisi
on
-
m
aking
keep
s
rising
i
n
m
ar
keting,
pa
rtic
ul
arly
with
the
adv
e
nt
of
th
e
In
te
r
net,
ra
pid
unde
rstan
ding
thr
ough
visu
al
rep
re
sentat
io
n
of
the
ef
fect
of
m
ark
et
ing
va
riables
on
str
at
egy
will
help
in
im
pro
ving
pro
fita
bili
ty
.
Sp
rea
dsheet
s
an
d
ot
her
non
-
vis
ual
dat
a
are
ve
ry
i
m
po
rta
nt
an
d
can
no
t
be
done
a
way
with.
It
is
best
t
o
prov
i
de
a
m
ark
et
in
g
a
naly
st
bo
t
h
visu
al
and
no
n
-
visu
al
data
s
o
that
s
ound
m
ark
et
ing
decisi
on
s
ca
n
be
m
ade.
S
om
e
m
a
nag
e
rs
are
best
at
un
der
sta
nding
nu
m
ber
s
an
d
oth
e
rs
are
m
os
tl
y
visu
al
;
he
nce
bo
t
h
m
us
t
be
pro
vid
e
d
to
m
anag
e
rs
f
or
m
akin
g
sound
de
ci
sion
s.
Heer
and
S
hnei
de
rm
an
[15]
sta
te
that
m
ulti
ple,
li
nk
e
d
vis
ualiz
at
ion
s
are
im
portant
f
or
pro
vid
in
g
m
eaningf
ul
insig
hts
into
m
ul
ti
di
m
ension
al
da
ta
rat
her than
is
olate
d
vi
su
al
iz
at
ion
of
t
he
sam
e
data
si
nce
t
he
q
ua
ntit
y
of
data
that
c
an
be
pr
ese
nted
i
n
a
sing
le
im
age
is
lim
it
ed
and
inter
-
relat
io
ns
hi
ps
be
twee
n
va
riables
an
d
data
set
s
ca
nnot
be
entirel
y
pr
ese
nt
ed
with
a
si
m
ple
i
m
age.
Effec
ti
ve
data
visu
al
isa
ti
on
an
d
unde
rstan
ding
the
aud
ie
nce
of
th
e
data
vis
ualisa
ti
on
is
cr
ucial
in
the
sal
es
e
nv
i
ronm
ent
as
it
al
lows
f
or
sal
es
per
s
onne
l
to
underst
and
th
e
internali
ze
t
he
visu
al
isa
ti
on
th
at
su
it
s
the
sal
es
pe
rs
onnel
st
yl
e
and
it
al
so
al
lows
t
he
ope
rati
on
al
pe
rson
nel
to
unde
rstan
d
the
internali
ze the
visu
al
isa
ti
on th
at
s
uits to t
heir
sty
le
.
Visu
al
iz
at
ion
un
ea
rths
to
pic
s
that
are
hid
de
n
due
to
the
c
om
plexity
of
the
is
su
e,
dri
ving
si
m
plific
at
ion
of
the
to
pic,
creati
ng
ur
gency
an
d
a
n
e
ffec
ti
ve
sen
se
of
the
opportu
ni
ty
cost
to
no
t
ta
ke
correct
ive act
ion
[1
1].Lastl
y, the i
m
ple
m
entat
ion
p
r
ocess o
f
a d
at
a
-
dri
ve
n
pro
j
ect
in
a sal
es env
ir
onm
ent
m
us
t
ens
ur
e
t
hat
ef
fecti
ve
data
vi
su
al
isa
ti
on
is
in
place
t
o
e
nsure
the
a
udie
nce
a
re
f
ully
eng
a
ge
d.
The
three
com
m
on
issues
that
m
us
t
alw
ay
s
be
ta
ke
n
into
co
ns
ide
rat
ion
are:
(i
)
Th
e
TIME
ta
k
en
or
data
gathe
ri
ng,
(ii)
The
in
div
id
ual
’s
ABIL
IT
Y
in
any
ind
ivid
ual
to
synthesiz
e,
analy
ze
under
s
ta
nd
the
data
vi
su
al
isa
ti
on
an
d
(iii
)
The
a
bili
ty
to
COMM
UNIC
ATE
t
he
ac
qu
i
red insig
hts t
o othe
rs wit
hin
t
heir
te
am
d
ow
n
the
li
ne.
ACKN
OWLE
DGME
NT
The
aut
hor
s
w
ou
l
d
li
ke
to
thank
Un
i
ver
sit
i
Teknolo
gi
Ma
la
ysi
a
fo
r
the
Po
te
ntial
Acad
em
ic
Staf
f
Gr
a
nt (Q.
K
130000.
2738.
03K
13).
REFERE
NCE
S
[1]
Singh
K,
W
aj
gi
R.
Da
ta
Anal
ysis
and
Visual
izati
on
of
Sal
es
Data
.
W
orld
Co
nfe
ren
c
e
on
Fut
uristi
c
Tre
nds
i
n
Resea
rch
and
In
nov
at
ion
for
Soc
ia
l
W
el
far
e, 201
6.
[2]
Bere
nson
M,
Levine
D,
Szabat.
Data
Visualiza
t
i
on
in
Marke
ti
ng
.
Journal
of
Marke
ti
ng
Mana
g
ement.
2015;
3
(2):
36
-
49.
[3]
Et
tor
chi
-
T
ard
y
A,
Le
v
if
M,
Mi
che
l
P.
Benc
hm
ark
ing:
A
Me
th
od
for
Conti
nuo
us
Quali
t
y
Im
pr
ovement
in
He
alth.
He
al
th
c
Poli
c
y
.
2012;
7
(4)
:
101
-
109.
[4]
Business
Dict
i
onar
y
,
"w
w
w.busine
ss
dic
ti
o
nar
y
.
com,"
[Onli
ne]
.
Available:
htt
p://ww
w.busine
ss
dic
ti
on
ar
y
.
c
om
/de
fini
ti
on
/metrics.ht
m
l. [Ac
c
essed
6
Ma
y
201
8]
.
[5]
Melonc
on
L
,
W
a
rne
r
E.
D
at
a
Vis
ual
i
za
t
ions:
A
Literature
Rev
ie
w a
nd
Opportunitie
s
for
Te
chnica
l
a
nd
Profess
iona
l
Com
m
unic
at
ion. Professional
Co
m
m
unic
at
ion
Co
nfe
ren
c
e
(ProCo
m
m
)
,
2017.
[6]
Kaluz
a
A,
Gell
r
ic
h
S,
Cerda
s
F
,
Thi
ed
e
S,
Herrm
ann
C.
Li
fe
C
y
c
le
Engi
n
ee
r
i
ng
Based
on
V
isual
Anal
y
tics.
Proce
dia CIRP
.
2018;
69:
37
-
42.
[7]
Saga
ert
YR,
A
ghez
z
af
EH,
K
oure
ntzes
N,
Desm
et
B.
Tactica
l
sal
es
fore
c
a
sting
using
a
ver
y
la
rg
e
set
o
f
m
ac
roe
conomic
indi
c
at
ors.
Euro
pea
n
Journa
l
of
Opera
ti
on
al
R
es
ea
rch
.
2018
;
264
(2):
558
-
56
9.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1340
–
1348
1348
[8]
Emeli
eHa
v
emo.
A
visual
per
spec
ti
ve
on
va
lue
c
r
ea
t
ion:
Expl
o
rin
g
pa
tt
ern
s
in
bus
ine
ss
m
odel
diagram
s.
Europe
a
n
Mana
gement
Jo
urna
l
.
2017.
[9]
Glea
nster
.
Gl
ea
n
sight
Ben
chmark
Report
-
Da
ta Visual
izat
ion. 201
5.
[10]
Ali
SM
,
Gupta
N,
Na
y
ak
GK
,
L
enka
RK.
Big
d
at
a
visualization
:
Tool
s
and
chal
le
nges.
Conte
m
p
ora
r
y
Com
puti
n
g
and
Inf
orm
at
i
cs
(IC3I).
2016
:
65
6
-
660.
[11]
Mage
e
B
,
Sam
m
on
D,
Nagle
T,
O’Ragh
al
l
ai
g
h
P.
Introdu
ci
n
g
data
driv
en
p
rac
t
ic
es
int
o
sa
l
es
envi
ronm
ent
s
:
exa
m
ini
ng
th
e
i
m
pac
t
of
da
ta
v
isual
isation
on
user
enga
g
ement
and
sa
le
s
resu
lt
s.
Journal
of
Dec
ision
S
y
s
tem
s
.
2016;
25
(1)
:
31
3
-
328.
[12]
Le
ppink
J.
Cogn
it
ive
lo
ad
the
or
y:
Prac
tical
impli
ca
t
ions
and
an
i
m
porta
nt
cha
l
le
n
ge.
Journal
of
T
ai
bah
Univer
si
t
y
Medic
a
l
Sci
ence
s
.
2017.
12
(5)
: 385
-
391.
[13]
Herna
ndo
A,
Bobadi
lla
J,
O
rte
ga
F,
Guti
ér
rez
A.
Method
to
in
te
ra
ctively
v
isual
i
ze
and
navi
ga
te
relat
ed
informati
on,
"
Ex
per
t
S
y
stems
wit
h
Applicati
ons,
2018.
[14]
Abela
.
Th
e
Ext
reme
Presenta
ti
on
(tm)
Method.
Septe
m
ber
2006.
[Online
]
.
Avail
able:
htt
p://ext
r
emepr
ese
ntation.t
y
pep
ad.
com/fi
le
s/
cho
osing
-
a
-
good
-
ch
art
-
09.
pdf
.
[A
ccess
ed
6
Ma
y
201
8]
.
[15]
Hee
r
J,
Shneid
er
m
an
B.
Inte
racti
ve
D
y
namics
for
Visual
Anal
y
sis
:
A
ta
xonom
y
of
tool
s
tha
t
support
the
flue
n
t
and
fle
xible
use
of
vi
suali
z
at
ions.
ACM
Queue
.
2016;
10
(2):
1
-
26
.
[16]
Jovanovic
J,
Ba
gher
i
E,
Gasevi
c
G.
Com
pre
hension
and
Le
arn
ing
of
Socia
l
Goals
through
Visuali
z
at
ion
.
IEEE
Tra
nsac
ti
ons on
Hum
an
-
Mac
hine S
y
stems
.
2015
: 478
-
489.
Evaluation Warning : The document was created with Spire.PDF for Python.