Statistika Rekayasa

Tabular & Graphical

Types of data
2
Categorical data (qualitative data)
Categorical Data
Use labels or names to identify an attribute of each
element.
Use either the nominal scale or ordinal scale of
measurement and may be nonnumeric or numeric.
The statistical analysis for qualitative data are rather
limited
Categorical variable is a variable with categorical
data
Examle!
Car tye "sedan# sort car# $U%# minivan# &'%# and so
on(
&ethod of ayment "cash# credit card# chec)(
*
Data set for 2 !utual fu"ds
+
#ua"titative data
,uantitative data
,uantitative data are always numeric.
Use either the interval or ratio scale of measurement.
-rdinary arithmetic oerations are meaningful only
with quantitative data.
,uantitative variable is a variable with
quantitative data.
Discrete if they are countable data and are collected
by counting. Ex! the number of items
Continuous if they are collected by measuring and
are exressed on a continuous scale. Ex! time to
failure of um comonent
.
#uestio"$ %hat&s the type of these data'
Tyes of shis
Categorical
/T$' students collecting data on the number
of shis entering inner channel of $urabaya
0est 1ccess Channel in a articular day.
#ua"titative discrete
Time until a fuel oil simlex filter getting
clogged u.
#ua"titative co"ti"uous
2
Tabular a"d graphical !ethods for
su!!ari(i"g data
3
Su!!ari(i"g categorical)qualitative data
4requency Distribution
5elative 4requency
'ercent 4requency Distribution
6ar 7rah
'ie Chart
8
*reque"cy distributio"
+ freque"cy distributio" is a tabular
summary of data showing the frequency "or
number( of items in each of several
nonoverlaing classes.
The ob9ective is to provide i"sights about the
data that cannot be quic)ly obtained by
loo)ing only at the original data.
:
,-a!ple$ .arada /""
;<
,-a!ple$ .arada /""
;;
Relative freque"cy distributio"
;2
0erce"t freque"cy distributio"
The perce"t freque"cy of a class is the
relative frequency multilied by ;<<.
1 perce"t freque"cy distributio" is a tabular
summary of a set of data showing the
ercent frequency for each class.
;*
,-a!ple$ .arada /""
;+
5elative frequency and ercent frequency distributions
1ar graph
1 bar graph is a grahical device for deicting
qualitative data that have been summari=ed in a
frequency# relative frequency# or ercent
frequency distribution.
-n the hori=ontal axis we secify the labels that
are used for each of the classes.
1 freque"cy# relative freque"cy# or perce"t
freque"cy scale can be used for the vertical axis.
Using a bar of fixed width drawn above each
class label# we extend the height aroriately.
The bars are separated to emhasi=e the fact
that each class is a searate category.
;.
1ar graph 2 .arada /""
;2
0ie chart
The pie chart is a commonly used grahical
device for resenting relative frequency
distributions for qualitative data.
4irst draw a circle> then use the relative
frequencies to subdivide the circle into
sectors that corresond to the relative
frequency for each class.
$ince there are *2< degrees in a circle# a class
with a relative frequency of <.2. would
consume <.2."*2<( ? :< degrees of the circle.
;3
0ie chart 2 .arada /""
;8
,-a!ple 2 .arada /""
@nsights 7ained from the
'receding 'ie Chart
-nehalf of the customers surveyed
gave &arada a quality rating of
Aabove averageB or AexcellentB
"loo)ing at the left side of the ie(.
This might lease the manager.
4or each customer who gave an
AexcellentB rating# there were t%o
customers who gave a AoorB rating
"loo)ing at the to of the ie(. This
should dislease the manager.
;:
,-a!ple 2 *ive popular softdri"ks
2<
Relative freq3 a"d perce"t freq3
distributio"s
2;
1ar chart a"d pie chart
22
Su!!ari(e of qua"titative data
4requency Distribution
5elative 4requency and 'ercent 4requency
Distributions
Dot 'lot
Cistogram
Cumulative Distributions
-give
2*
,-a!ple$ 4udso" +uto Repair
2+
The manager of Cudson 1uto would li)e to get a better
icture of the distribution of costs for engine tuneu arts. 1
samle of .< customer invoices has been ta)en and the costs
of arts# rounded to the nearest dollar# are listed below.
*reque"cy distributio"
7uidelines for $electing Dumber of Classes
Use between . and 2< classes.
1roximate formula to calculate number of
class may also be introduced as!
ra"ge 5 largest data value 2 s!allest data value
class 5 k 5 67838 log 9: %here 9 is "u!ber of sa!ples
i"terval 5 ra"ge)class
Data sets with a larger number of elements usually
require a larger number of classes.
$maller data sets usually require fewer classes.
2.
*reque"cy distributio"
7uidelines for selecting width of classes
Use classes of equal width.
22
Approximate class width =
Largcst data vaIuc -SmaIIcst data vaIuc
Numbcr oI cIasscs
,-a!ple$ 4udso" +uto Repair
23
4requency distribution
,-a!ple$ 4udso" +uto Repair
28
,-a!ple$ 4udso" +uto Repair
2:
@nsights gained from the ercent frequency
distribution!
-nly +E of the arts costs are in the F.<.: class.
*<E of the arts costs are under F3<.
The greatest ercentage "*2E or almost onethird( of the
arts costs are in the F3<3: class.
;<E of the arts costs are F;<< or more.
Dot plot
-ne of the simlest grahical summaries of
data is a dot lot.
1 hori=ontal axis shows the range of data
values.
Then each data value is reresented by a dot
laced above the axis.
*<
,-a!ple$ 4udso" +uto Repair
*;
4istogra!
1nother common grahical resentation of
quantitative data is a histogram.
The variable of interest is laced on the
hori=ontal axis and the frequency# relative
frequency# or ercent frequency is laced on the
vertical axis.
1 rectangle is drawn above each class interval
with its height corresonding to the intervalGs
frequency# relative frequency# or ercent
frequency.
Unli)e a bar grah# a histogram has no natural
searation between rectangles of ad9acent
classes.
*2
,-a!ple$ 4udso" +uto Repair
**
Cu!ulative distributio"
The cumulative frequency distribution shows
the number of items with values less than or
equal to the uer limit of each class.
The cumulative relative frequency
distribution shows the roortion of items
with values less than or equal to the uer
limit of each class.
The cumulative ercent frequency
distribution shows the ercentage of items
with values less than or equal to the uer
limit of each class.
*+
,-a!ple$ 4udso" +uto Repair
*.
;give
1n ogive is a graph of a cu!ulative distributio".
The data values are shown on the hori=ontal
axis.
$hown on the vertical axis are the!
cumulative frequencies# or
cumulative relative frequencies# or
cumulative ercent frequencies
The frequency "one of the above( of each class is
lotted as a oint.
The lotted oints are connected by straight
lines.
*2
,-a!ple$ 4udso" +uto Repair
-give
6ecause the class limits for the artscost data
are .<.:# 2<2:# and so on# there aear to be
one unit gas from .: to 2<# 2: to 3<# and so on.
These gas are eliminated by lotting oints
halfway between the class limits.
Thus# .:.. is used for the .<.: class# 2:.. is used
for the 2<2: class# and so on.
*3
,-a!ple$ 4udso" +uto Repair
*8
-give with Cumulative 'ercent 4requencies
,-ploratory data a"alysis
The techniques of exloratory data analysis
consist of simle arithmetic and easytodraw
ictures that can be used to summari=e data
quic)ly.
-ne such technique is the stemandleaf
dislay.
*:
Ste!<a"d<leaf display
1 stemandleaf dislay shows both the ran)
order and shae of the distribution of the data.
@t is similar to a histogram on its side# but it has
the advantage of showing the actual data
values.
The first digits of each data item are arranged to
the left of a vertical line.
To the right of the vertical line we record the last
digit for each item in ran) order.
Each line in the dislay is referred to as a stem.
Each digit on a stem is a leaf.
+<
,-a!ple$ 4udso" +uto Repair
+;
Stretched ste!<a"d<leaf display
@f we believe the original stemandleaf
dislay has condensed the data too much# we
can stretch the dislay by using two more
stems for each leading digit"s(.
0henever a stem value is stated twice# the
first value corresonds to leaf values of <+#
and the second values corresonds to values
of .:.
+2
,-a!ple$ 4udso" +uto Repair
+*
Ste!<a"d<leaf display
Heaf Units
1 single digit is used to define each leaf.
@n the receding examle# the leaf unit was ;.
Heaf units may be ;<<# ;<# ;# <.;# and so on.
0here the leaf unit is not shown# it is assumed to
equal ;.
++
,-a!ple$ =eaf u"it 5 >36
+.
,-a!ple$ =eaf u"it 5 6>
+2
Crosstabulatio"s a"d scatter diagra!s
Thus far we have focused on methods that
are used to summari=e the data for one
variable at a time.
-ften a manager is interested in tabular and
grahical methods that will hel understand
the relationshi between two variables.
Crosstabulation and a scatter diagram are
two methods for summari=ing the data for
two "or more( variables simultaneously.
+3
Crosstabulatio"
Crosstabulation is a tabular method for
summari=ing the data for two variables
simultaneously.
Crosstabulation can be used when!
-ne variable is qualitative and the other is
quantitative
6oth variables are qualitative
6oth variables are quantitative
The left and to margin labels define the
classes for the two variables.
+8
,-a!ple$ *i"ger =akes 4o!es
+:
Crosstabulation
The number of 4inger Ha)es homes sold for each style
and rice for the ast two years is shown below.
,-a!ple$ *i"ger =akes 4o!es
.<
@nsights 7ained from the 'receding
Crosstabulation
The greatest number of homes in the samle ";:( are
a slitlevel style and riced at less than or equal to
F::#<<<.
-nly three homes in the samle are an 14rame style
and riced at more than F::#<<<.
Crosstabulatio"$ Ro% or colu!"
perce"tages
Converting the entries in the table into row
ercentages or column ercentages can
rovide additional insight about the
relationshi between the two variables.
.;
,-a!ple$ *i"ger =akes 4o!es
.2
Scatter diagra!
1 scatter diagram is a grahical resentation
of the relationshi between two quantitative
variables.
-ne variable is shown on the hori=ontal axis
and the other variable is shown on the
vertical axis.
The general attern of the lotted oints
suggests the overall relationshi between the
variables.
.*
Scatter diagra!
.+
,-a!ple$ 0a"thers *ootball Tea!
..
,-a!ple$ 0a"thers *ootball Tea!
.2
,-a!ple$ 0a"thers *ootball Tea!
The receding scatter diagram indicates a
ositive relationshi between the number of
intercetions and the number of oints
scored.
Cigher oints scored are associated with a
higher number of intercetions.
The relationshi is not erfect> all lotted
oints in the scatter diagram are not on a
straight line.
.3
Refere"ces
$tatistics for 6usiness and Economics# 1nderson#
$weeney# and 0illiams# 0est 'ublishing Comany.
$tatistics for 6usiness and Economics.#
$outh0esternIThomson Hearning
.8