FACTOR ANALYSIS

PL5101 | Planning Analytical Method

Group 3 Group 8

Fathurrahman 25422005 Bakhtiar Luthfi Fakhrozi 25422002

Dita Mulyani 25422022 Dhea Ananda Putri 25422019
Bintang Putri Mahayana 25422039 Ahmad Baikuni Perdana 25422036
 TODAY’S MENU
1. To understand the definition of Factor Analysis.

2. To understand the concept of Factor Analysis.

3. To understand the function of Factor Analysis.

4. To know the approach types of Factor Analysis.

5. To know the assumptions of Factor Analysis.

6. To know the application of Factor Analysis

in Urban and Regional Planning.
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Source: Dillon,William R. 1984. Multivariate Analysis Methods and Applications, p.54

Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

• Data reduction technique for investigating

interdependences .
• The study of interrelationships among the
variables in an effort to find a new set of
variables, fewer in number than the
original set of variables.
• Multivariate Analysis. To find which is common among the
is a technique of analyzing a
large number of object original variables.
variables in one or more
samples simultaneously. “Factors” or “Underlying
Dimensions” refers to factors that
only represent common or shared
variation.
Source: Dillon,William R. 1984. Multivariate Analysis Methods and Applications, p.54
Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Factor analysis is statistical technique applied to a single set of variables when the researcher is
interested in discovering which variable in the set form coherent subset that are relatively
independent of one and another (Tabachnick: 2007).

Waiting Time
Say that you want to pick
Cleanliness Service
a restaurant to go to, so
you start checking Staff behavior
reviews. You start to
categorize the aspect of Taste of the food

six variables which are Food freshness Food quality

Food temperature

Source: Tabachnick BG dan Fidell LS, Using Multivariate Statistics, Fifth Edition, Pearson, 2007
Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

X1
X2 X6
X3 X9
X5
X4 X8
X7

Nine correlated variables

How do you define the common factors or underlying dimensions

among the 9 variables?

Source: Dillon,William R. 1984. Multivariate Analysis Methods and Applications, p.54

Factor Analysis - Variable to Factor Reduction
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Nine variables reduced to three factors

X1
X3 X2 X5
X2 X6
X3 X9 X1 X4
X5 X8 X9
X6 X7
X4 X8
X7

Factor 1 Factor 2 Factor 3

Nine correlated variables

How do you define the common factors or underlying dimensions

among the 9 variables?
Source: Dillon,William R. 1984. Multivariate Analysis Methods and Applications, p.54
Factor Analysis - Variable to Factor Reduction
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Nine variables reduced to three factors

Instead of having to deal with the

nine variables separately, we now X5
need to consider only the three X3 X2
factors as defined in the figure
(Dillon: 1984, p.54). X1 X4 X8 X9
X6 X7
So, there are two types of variable.
The nine correlated variables are the
observed variables while the Factor
1, 2, and 3 are the latent variables Factor 1 Factor 2 Factor 3
or derived variables.
Warm Tones Patterns Cool Tones

Source: Dillon,William R. 1984. Multivariate Analysis Methods and Applications, p.54

Function of Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Summary of Data Identification of

This feature allows us to use a Underlying Factors
few factors, perhaps only one By clustering a large number of
or two, to account for the bulk variables into a smaller number
of the variance contained in the of homogeneous sets and
entire set of variables. creating a new variable a
factor representing each of
these sets

Screening of Variables Sampling of Variables

factor analysis identifies Selecting a small group of
groupings of variables that are uncorrelated variables among
highly correlated with one a large set of variables
another, we can choose a explains practical problems
single variable from each
factor, thereby avoiding the
collinearity problem.
Latent Variable
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Age
Latent variables are
variables that are not Gender Morals
directly observed but rather Morals
Experience
inferred from other
variables that are observed.
This is what we call factors Income
its actually difficult to
Social class Confidence
measure numerically
Education
Examples of the latent
Happiness Confidence
variables are quality of life,
Income
business confidence,
morals, and happiness Married Happiness

Age
 Certain Assumptions are made about the data
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

The sample size is supposed Matric variables are

to be a greater function expected (interval data)
in general principles is the The variables are expected
minimum 1 to 5 ratio which to be of numeric type, they
is 10 factors you should have should be in intervals of

1 around 50 samples at least

3 numbers
5

There is no outliner
in the data
2 Variables must be
interrelated
4 Multivariate
normality is not
We assume that our data is (Barret test) required
clean, there should be no we can perform something
outliers or missing values called a buried test to analyze
the correlation
 Step of Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Formulation Indicator Correlation Factor Factor

Problem Matrix Extraction Rotation
Purpose Kaiser Mayer Olkin • Principal • Varimax method Factor
(formulating goals (KMO) Measure of Component Minimize the Interpretation
using factor sampling adequacy is Analysis number of
analysis), Amount an index used to • Principal axis variables with
of data(measuring examine the factoring extreme loading
scale), and appropriateness of • Unweight least (high or low) on a
Number of factor analysis. squares factor.
samples (the Bartlett’s test of • Generalized last • Quatimax method
sample used is sphericity examines square • Equamax method
between 50-100 the hypostasis that • Maximum likelihood
the variables are
with a comparison
uncorrelated in the
scale of 1:5) population

KMO > 0,5

Ratio 1:5
SIG < 0,05 choose one
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Input Data
In input data for processing factor analysis, there are two
types of data that can be input:

Covariance: the existing data have the same size/unit/scale

Correlation: the existing data do not have the same size/unit/scale

If there is a form of correlation data, where the data are not in the same size/unit/scale,
then standardization of the size/unit/scale can be done. Standardization to equalize
units is done based on the standard deviation value
 Step of Factor Analysis
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Naming the Factors

The process involves substantive interpretation of the pattern of factor loadings for the variables,
including their signs, in an effort to name each of the factors. Before interpretation, a minimum
acceptable level of significance for factor loadings must be selected. Then, all significant factor
loadings typically are used in the interpretation process. Variables with higher loadings influence to a
greater extent the name or label selected to represent a factor.

Multivariate Data Analysis by Jr., William C. Black, Barry J. Ba Joseph F. Hair,2013

 Types of factor analysis approach
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

EFA : used to reduce data to a smaller set of summary variables and to explore the underlying
theoretical structure of the phenomena.

CFA : used to test how well the measured variables represent the number of constructs

EFA CFA
(exploratory factor analysis) (confirmatory factor analysis)
Basic Assumptions Total variance (unique, error, common) is Only common variance is taken into
taken into account to drive the factor. consideration

Purpose Is to discover the underlying structure in the data Assume the existence of latent factors
using a correlation matrix, it is used for getting underlying the observed data
insight out of the data

Factor selection When they search for latent patterns in the data A priori based on known theories and
case there is no or only little prior knowledge on concepts.
how variables are related
Main Technique Principal Component Analysis Principal Factor
Variancenalysis (PCA)
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Variance is Average squared deviations of each score from the mean of the scores.

Respondent Weights Mean Deviations Squared

s (Kg) (Variable Deviations
-Mean)

A 30 35 -5 25

B 35 35 0 0

C 40 35 5 25

● Average Squared Deviations = (25+0+25)/(3-1) =

25
● Standard Deviations=squared of variance = 5
Variancenalysis (PCA)
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Common Variance &

Unique Varians Unique Variance
It has no connection to any
other factors.

Common Variance
have a connection between
the variables

Component Factor
Analysis
Common Factor Analysis
are all extracted without
the analysis observed on
regard to common
common variance and
variance, and do not pay
unique variance
attention to overlap
Total VarianceAnalysis (PCA)
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

Communality is the sum of squared multiple Communality of “COST”=

correlation of variables as predicted from the (-0.086)^2+(0.981)^2=0.970
factors Total Variance=1-0.970=0.03
Principal ComponentAnalysis (PCA)
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

● PCA is an approach that tests the total variance

in the data.
● It is empirical, there is no theory that underlies
the formation of factors before the analysis is
carried out

PC = A.F + U+e

A = Loading Matrix
F = Common Factor Matrix
U = Unique Factor Matrix
e = Error Value of x
Common Factor (CF)Analysis (PCA)
OUTLINE DEFINITION CONCEPT FUNCTION ASSUMPTION APPROACH

● CF is an approach that distinguishes common and unique variance, and only tests
common data.
● Based on theory and is an approach that serves to confirm the hypothesis.

● X1 = v11CF1 + v12CF2 +……..+ v1mCFm +

e1
● X2 = v21CF2 + v22CF2 +……..+ v2mCFm +
e2
● Xp = vp1CFp + vp2CF2 +……..+ vpmCFm +
ep
APPLICATION
OF FACTOR ANALYSIS
In The Context Of Urban and Regional Planning
Application Of Factor Analysis
In The Context Of Regional And City Planning
"Factors Influencing the Selection of Tlogosari Housing Areas
as Residents of Semarang City".

Formulation of the problem:

● What variables have the most influence on the selection of residential
areas
● What variables can be reduced & grouped as Factors (a new collection
of variables) that affect a person's preferences in choosing the Tlogosari
Housing Area as a place to live for the residents of Semarang City
● How are new Factors formed?

Research purposes :
● Knowing the influencing factors
● Knowing the new set of variables (Factors) formed from the variable
Selection of the Tlogosari Housing Area as a Place to Live for Semarang
City Residents
Variable Data
The preparation of research variables is exploratory or there is no theory/hypothesis
that composes the structure of the formed factors.
Variables in determining the location of settlements according to Bourne (1978):

Accessibility to city center

Physical characteristics of the

residential environment

Facilities and services

Social environment

Site and house characteristics

 Process of SPSS
1 2

See The “Variable View” to

determine the variable’s name

Add Data Into “Data View”

 Process of SPSS
3 4

After we choose determination

reduction and factor analysis, we will
see this page

Click Analyze and

Factor analysis
 Process of SPSS
5 6

We move the variable

Click “Descriptive” and

check as shown in the
picture
 Process of SPSS
7 8

Click “Extraction” and check as Click “Rotation” and check as

shown in the picture shown in the picture
 Process of SPSS
10
9

Click “Scores” and check as Click “Option” and check

shown in the picture as shown in the picture
 Final Result of SPSS

11 12

After we change the settings as

needed, then we click “ok”

After clicking OK, the results of the

factor analysis will appear
Interpretation of Analysis Result

Accessibility to the city center is the

variable with the largest mean value,
as can be seen in the descriptive
statistics table.
Interpretation of Analysis Result
• The correlation table explains the magnitude of the correlation between the variables to be analyzed. The closer to 1, the
greater the correlation value.
• The sig.(1-tailed) shows the significance of the correlation between these variables. If the p-value (Sig.) < 0.05 then there is a
correlation.

There is a significant correlation:

● Characteristics of environment among
Accessibility to City Center, Social and
Characteristics of Site and house
● Accessibility to City Center among
Characteristics of environment, Social
and Characteristics of Site and house
● Social among Characteristics of
environment, Accessibility to City
Center and Characteristics of Site and
house
● Characteristics of Site and house
among Characteristics of
environment, Accessibility to City
Center and social
Interpretation of Analysis Result

This KMO method measures the overall sampling

and measures the sampling evaluation for each
indicator.

It is possible to conduct an analysis because the

KMO value of 0.711 indicates that the p-value is >
0.5.

PCA is possible because the significance value (Sig.)

on Bartlett's Test of Sphericity is 0.000 > 0.05.
Interpretation of Analysis Result

image matrices to determine whether

the variables are partially feasible to
be analyzed and not excluded from
testing if MSA>0.5.

There are variables that cannot be

predicted and cannot be analyzed
further or excluded from other
variables because the MSA value of
these variables is less than 0.5.
Variables are excluded starting from
the smallest variable.
Interpretation of Analysis Result

Based on the scree plot, there is 1 new

factor formed, seen from the eigenvalue
of more than 1.
Interpretation of Analysis Result

•If the results of the component matrix, grouping ● One factor formed
variables into factors, are still difficult to decide, then ● Naming is done based on the variable with
rotation can be done. the largest loading factor value from each
•The – or + sign only shows the direction of the
correlation, so it is still difficult to group ● The name of the factor is Social
CASE STUDY Using EFA
Case Study Outlines
ANALYSIS OF CITIZENS' MOTIVATION AND PARTICIPATION INTENTION IN URBAN
PLANNING

Argument :
There are very few theory-driven empirical studies in the participation literature that examine citizens’
motivation and intention to participate in the urban planning process.

Purpose :
To fill this gap in the literature by investigating the relationship between motivation and the intention of
public participation in the field of urban planning and management.

Basis :
A survey conducted in Wuhan, China; developed into a structural equation model
● Purpose : To identify the causal relationships between four motivational factors and three
intentions to participate in urban planning.
● Defined factors : ‘Civil Society, ‘Personal Interest’, ‘Social Influence’, and ‘Constraints’
● Measured intentions: ‘Urban Planning Scale’, ‘Urban Planning Content’, and ‘Urban Planning
Process
The Research Process
Latent Variable in Urban Participatory Planning

Civil Society
Intrinsic
factor
Urban Planning
Motivation Factors

Personal Interest Scale

Urban Planning
Urban Planning
Participate
Content
Intention
Social Influence Urban Planning
Process
Extrinsic
factor
Constraints
 Variable Constructs
CONSTRUCTION
EFA Results
1. Goal:
● find out the major motivational factors
and intention factors of urban planning
participation
● extract the essential structure of
multivariate observation variables and
process dimensionality reduction
2. KMO Test: 0.918
3. Total Explained Variance: 72.9%, good
reliability of the scale
4. All Observed Variables’ coefficients greater
than 0.7. It indicates they are sufficient to
represent and interpret the latent variable
Study Case Conclusion

Civil Society

Urban
Personal Planning Scale
Interest
Urban
Planning
Content
Social Urban
Influence Planning
Process
Constraints
Lesson Learned
Factor Analysis can be used to identify the major
1 motivational factors and intention factors of urban
planning participation in urban planning (Li, et. al, 2020)

Factor Analysis’ result is not always reducing the

2
variables.

We can select the number of factors by the theoretical

3 processes underlying a research area (Tabachnick B.G
& Fidell L.S, 2007).

We need to know the difference between

4 Confirmatory Factor Analysis and Exploratory
Factor Analysis.
SUMMARY
OF FACTOR ANALYSIS
Summary
of Factor Factor Analysis (FA) can be used to:

Analysis • Summary of Data

• Screening of Variables
• Identification of Underlying Factors
• Sampling of Variables
There are two types of FA:
• Confirmatory FA
• Exploratory FA
Implementation of FA in Urban and Regional Planning:
• Participation Intention in Urban
Planning;
• Factors in housing location choices;
• etc
Reference

❑ Dillon, W dan Goldstein M, Multivariate Analysis: Methods and Aplication, John Willey & Son, 1984
❑ Tabachnick BG dan Fidell LS, Using Multivariate Statistics, Fifth Edition, Pearson, 2007
❑ K, Stastical Analysis: Introduction to Bivariate & Multivariate Analysis , Radius Press ,
1986.(PustakaPendukung)
❑ Jr., William C. Black, Barry J. Ba Joseph F. Hair, Multivariate Data Analysis, Seven Edition ,2013
❑ Li, W., Feng, T., Timmermans, H. J. P., Li, Z., Zhang, M., & Li, B. (2020). Analysis of citizens' motivation and
participation intention in urban planning. Cities, 106, [102921].
https://doi.org/10.1016/j.cities.2020.102921
FACTOR ANALYSIS
Thank
You