SOCIAL NETWORK ANALYSIS II

MTP
ISMU RINI DWI ARI
JURUSAN PERENCANAAN WILAYAH DAN KOTA FTUB
MATERIALS

I. CONCEPT OF CENTRALITY
II. EXAMPLE OF CENTRALITY
III. AFFILIATION DATA
IV. CENTRALITY
V. RATE OF PARTICIPATION
I. CONCEPT OF CENTRALITY (SCOTT J, 1991)

 is a measure of the importance of a node in providing

connectivity (bridging) to other nodes in the network.

 is a measure of the contribution of network position to the

importance, influence, prominence of an actor in a network.
I. CONCEPT OF CENTRALITY (SCOTT J, 1991)

 Local centrality (LC) of a node

 Absolute LC is simply the degree of a node
 Relative LC is ratio between the degree of connection to the total number of nodes
exclude the node of itself

 Global centrality a node

 is defined as the sum of all distances of that node to the rest of the nodes of the
network
II. EXAMPLE OF CENTRALITY (SCOTT J, 1991)
E F J K N

A G B C O
D M

H L Q P

A,C B G,M J,K,L All other nodes

Local Absolute 5 5 2 1 1
Centrality
Relative 0.33 0.33 0.13 0.07 0.07
Global Centrality 43 33 37 48 57

 A node is globally central if it lies at short distances from the other nodes
of the network.
III. AFFILIATION NETWORK

 The importance of studying affiliation networks is grounded in

theoretical importance of individuals’ memberships in collectivities.

 Connections among members of one mode of the modes are based

on linkages established through the second mode.

Source: Wasserman & Faust, 1994

III. AFFILIATION NETWORK

 Such network of actors tied to each other through their participation in

collectivities, and collectivities linked through multiple memberships of actors.

 Consist of two elements:

 A set of actors and
 A collection of subsets of actors (events)

Source: Faust, 1997

III. AFFILIATION NETWORK
 The set of actors is labeled as N  {n1 , n2 ,..., ng }
The set of events is labeled as M  {m1 , m2 ,..., mh }
 Affiliation Matrix A  aik   gxh
m1 m2 m3 m4
n1 1 0 1 1
n2 0 1 0 1
n3 0 1 1 0
n4 0 0 1 0
 Hypothetical Example:
n5 1 1 0 1
6 actors & 4 events n6 1 1 1 1

A ‘1’ in row actor i , column event ofkA indicates that actor ni is affiliated
8
with event mk
III. AFFILIATION NETWORK : ONE-MODE

n1 n2 n3 n4 n5 n6
n1 0 1 1 1 2 3

One-mode matrix of actor n2 1 0 1 0 2 2

co-memberships – indicates n3 1 1 0 1 1 2
the number of memberships
n4 1 0 1 0 0 1
shared by each pair of actors
n5 2 2 1 0 0 3
n6 3 2 2 1 3 0
X  AA'
N

9
III. AFFILIATION NETWORK : TWO- MODE

n1 n2 n3 n4 n5 n6 m1 m2 m3 m4
n1 0 0 0 0 0 0 1 0 1 1
n2 0 0 0 0 0 0 0 1 0 1 Bipartite Matrix
n3 0 0 0 0 0 0 0 1 1 0
0 A
NM

0 
n4 0 0 0 0 0 0 0 0 1 0 X
n5 0 0 0 0 0 0 1 1 0 1  A'
n6 0 0 0 0 0 0 1 1 1 1
m1 1 0 0 0 1 1 0 0 0 0
m2 0 1 1 0 1 1 0 0 0 0
m3 1 0 1 1 0 1 0 0 0 0
m4 1 1 0 0 1 1 0 0 0 0
10
IV. CENTRALITY FOR AFFILIATION NETWORKS
 The notion of centrality is to identify the most important actors in a
social network that located in strategic locations within the network
(Wasserman and Faust,1994).

 The measure of the prominence or importance of the actors in a social

network.
 A prominent actor is if the ties of the actor make the actor particularly
visible to the other actors in the network.
 Based on the relational pattern of the row and column entries.

 The use of centrality measures is to understand group structure

which those actors with the most access will be the most central in
11
the network.
IV. CENTRALITY FOR AFFILIATION NETWORKS
ONE-MODE DEGREE CENTRALITY (WASSERMAN & FAUST, 1994)
 The simplest definition of actor centrality is that central actors must be the most
active in the sense that they have the most ties to other actors in the networks.
 This very active actor should thus have a maximal centrality index.
 The most visible actors in the network.

 Index of actor degree centrality

C D (ni )  d (ni )  xi    xij   x ji

j j
 Normalized degree centrality

d (ni )
C ' D (ni ) 
g 1

 An ego density for a non-directional relation is simply the ratio of degree of an actor to
the maximum number of ties that could occur.
IV. CENTRALITY FOR AFFILIATION NETWORKS
TWO-MODE DEGREE CENTRALITY

 Affiliation network as two-mode networks, consists of subsets of actors

wherein connections among members of one of the modes are based on
linkages established through the second mode (Wasserman & Faust, 1994).

 Two-mode Degree Centrality

 The actors are important because of their level of activity or the number of
contact that they have (Faust, 1997).
 The raw degree centrality for an actor is defined as the number of events that is
participated by the actor
CD (ni )

Actor ni normalized degree centrality C D ( ni )


h
(Everett & Borgatti, 2005)
IV. CENTRALITY FOR AFFILIATION NETWORKS
ONE-MODE CLOSENESS CENTRALITY (WASSERMAN & FAUST, 1994)
 Closeness measures how close an actor is to all the other actors in the set of actors as a
function of geodesic distance.

 Actors are central if they can contact others through efficient

(short) paths
 Index of actor closeness centrality: Sabidussi’s (1966) in Wasserman & Faust
(1994)
1
 g
CC (ni )   d (ni ,n j )
 C 'C (ni )  g 1
 ( g  1)CC (ni )
 j 1 

 d ( ni , n j ) 
g
 j 1 

 Normalized closeness centrality: Beauchamp (1965) in Wasserman & Faust

(1994)
d ( ni , n j ) : the length of any shortest path between the actors for all j ≠ i
14

d (ni , nk )   : the geodesic distance of the isolate actor nk = 0 for all i ≠ k

IV. CENTRALITY FOR AFFILIATION NETWORKS
ONE-MODE BETWENNESS CENTRALITY (WASSERMAN & FAUST, 1994)

 is that an actor is central if it lies between other actors on their geodesics,

implying that to have a large “betweenness” centrality, the actor must be
between many of the actors via their geodesic.
 Index of actor betwenness for ni is the sum of the estimated probabilities
over all pairs of actors not including the ith actor for i distinct from j and k.

C B (ni )   g jk (ni ) / g jk
j k
 Normalized betwenness centrality

C 'B (ni )  CB (ni ) /[( g  1)( g  2) / 2]

 The value is between 0 and 1

g jk : the number of geodesic linking the two actors

g jk (ni ) : the number of geodesic linking the two actors that contain actor i 15

( g  1)( g  2) / 2 : the number of pairs of actors not including ni

PARTICIPATION APPROACH ON COMMUNITY BASED
WATER MANAGEMENT
CASE STUDY I

ISMU RINI DWI ARI

DISSERTATION – KYOTO UNIVERSITY
MEMBERSHIPS IN COMMUNITY GROUP

Religious Cultural/Social Group

 MEMBERSHIP IN COMMUNITY GROUP

Community Org. Finance

 (TWO-MODE) DEGREE CENTRALITY
Sumberawan Ngujung RW 6
Mean 0.300 0.320 0.232
Min 0 0 0
Max 0.750 0.750 0.750
Std. Dev. 0.163 0.169 0.172
Variance 0.027 0.029 0.030
Level of Centrality (no. of respondents)
0 – 0.333 71 42 58
0.334 – 0.666 28 20 11
0.667 – 1 1 2 1

Sumberawan Ngujung RW 6
Name Mr. Suwandi Mrs. Jumiati Mrs. Suyati Mr. Supomo
19
Occupation Head of hamlet Small private business Farmer
Secretary of HIPPAM in Singosari
CLOSENESS (C’C) & BETWEENNESS (C’B) CENTRALITY
Sumberawan Ngujung RW 6
Centrality C’C C’B C’C C’B C’C C’B
Mean 0.956 0.0007 1 0 0.858 0.0041
Min 0.530 0 1 0 0.505 0
Max 1 0.0065 1 0 0.981 0.0708
Std. Dev. 0.080 0.0014 0 0 0.125 0.0124
Variance 0.006 0 0 0 0.016 0.0002
Level of Centrality (no. of respondents)
0 – 0.333 0 89 0 0 0 52
0.334 – 0.666 3 0 0 0 6 0
0.667 – 1 86 0 58 0 46 0

Almost each respondent has direct contact with every other respondents. 20

None of respondent that occupies position as a mediatory between others.

 SUMMARY

Hamlet Centrality Rate of

Degree Closeness Betweenness Participation

Sumberawan Strong power Middle & high Almost no mediator at least participate
relation in one community
No power relation High No mediator group
Ngujung
RW 6 Low power Middle & high Almost no mediator
relation

 Sumberawan: single choice of piped clean water supply system from HIPPAM is a result of their
collective action driven by central leader since in general every respondent may has high equal
influence to each other within the community.
 Ngujung & RW 6: choices upon piped clean water supply system is a result of an individual action,
whereby among residents they might have quite high connectedness and equal ability to influence each 21

other, but there is no major power relation that able derived them to the similar option.
CLIMATE CHANGE, FLOODS AND HOMES:
A SOCIAL NETWORKS APPROACH TO UNDERSTANDING
LOCATION PREFERENCES IN INDONESIA
CASE STUDY II

Asian Cities Climate Resilience, IIED, ACCRN

Working Paper Series 17: 2015 MUSTIKA ANGGRAENI, ISMU RINI DWI ARI AND ENDRATNO BUDI SANTOSA
PETA SITUASI

SUNGAI BENGAWAN SOLO

TANGGUL DIBANGUN OLEH

PEMERINTAH

DAERAH BANJIR

1,100 RUMAH & SEKITAR 1,300

KELUARGA TINGGAL DI LUAR
TANGGUL
BANJIR DAN DAMPAKNYA
50
45
40
35
30
25
20
15
10
5
0
1-3 times 3-5 times 5-7 times >7 times
Flooding Disarter Frequencies per Year

Dalam mengantisipasi kejadian banjir, sebuah

tanggul dibangun sepanjang sungai Bengawan
Solo pada wilayah studi. Ketinggian tanggul
bervariasi antara 1 – 1.5 meter. Fungsi tanggul
menjadi sangat buruk karena penduduk
membuka tanggul sebagai akses dari area tidak
terlindung ke pusat kota yang terlindungi.
Fungsi yang buruk ini pada dasarnya terkait
dengan perilaku penduduk.
SKEMA PENELITIAN
PERTANYAAN PENELITIAN

Bagaimana struktur sosial komunitas di area rawan banjir? Mengapa penduduk memilih lokasi
bermukim di area rawan banjir?

HIPOTESA

Rumah tangga dengan ikatan sosial yang lebih baik mempunyai pengaruh lebih besar dalam penentuan
lokasi rumah di area rawan banjir

TUJUAN

Menyelidiki mekanisme penentuan lokasi rumah tinggal bagi penduduk di area rawan banjir
NETDRAW – HIGH VULNERABILITY COMMUNITY

RT 2 RT 1

RT 3
JETAK RW 1 ( RT 1, 2, 3)
NETDRAW – MEDIUM VULNERABILITY COMMUNITY

RT 4
JETAK RW 1 (RT 4, 5)

RT 5
NETDRAW – LOW VULNERABILITY COMMUNITY

RW 1 (RT1, 2, 5)
RW 3 (RT 22)

RW 2 (RT 9)

BANJARJO RW 1 (RT1, 2, 5), RW 2 (RT 9) , & RW 3 (RT 22)

V. RATE OF PARTICIPATION

 Describe people’s rates of participation in social activities

(McPherson, 1982).

 The mean rate of affiliation for actors in the bipartite matrix that
might one to compare people’s rate of participation in voluntary
organization between communities (Wasserman and Faust,
1994).

Source: Wasserman and Faust (1994)

(I). RATE OF PARTICIPATION

 The row totals of an affiliation matrix, A or

 The entries on the main diagonal of actor co-memberships matrix, X N in a
valued relation matrix.
 The number of events with which actors is affiliated is denoted by
ai    j 1 aij  xiiN
h

h = the number or the size of events, M , to which actor belongs.

i 1  j 1 aij i 1 ii
g h g N
a  x
ai   
g g g
Source: Wasserman and Faust (1994)
(II) SIZE OF EVENTS

 The average number of actors in each event, or mean degree of events in

the bipartite matrix.
 The size of each event refers to the number of actors affiliated with event j
is given by the column total of affiliation matrix, A, or the entries on the
main diagonal of the event overlap matrix, X M, is denoted by

a j  i 1 aij  x
g M
jj

i 1  j 1 aij  j 1 jj
g h h M
a  x
a j   
h h h
Source: Wasserman and Faust (1994)
TERIMA KASIH

 Faust, K., 1997, Centrality in Affiliation Networks, Social Networks 19:157-191

 Sabidussi, G., 1966, The Centrality Index of a Graph, Psychometrika Vol. 31 No. 4.
 Scott, J., 1991, Social Network Analysis: A Handbook, 2nd Edition, SAGE.
 Wasserman, S. and Faust, K., 1994, Social Network Analysis: Methods and Applications,
Cambridge University Press.