Business Network

Analytics

Sep 12th, 2017

Daning Hu
Department of Informatics
University of Zurich
Business Intelligence
Research Group

F Schweitzer et al. Science 2009

 Business Network Analytics and Business
Intelligence
Economic
Networks

Financial Markets, Monitoring, Analyzing, and

Risk Contagion
Banking Systems… Simulating Contagious
(e.g., Bank run)
Risks, …

Network Modeling Design

Influence
& Analysis BI Applications

Online Communities, Social Contagion Identification and

Consumer Social (e.g., Word-of- Prediction of Influence-
Networks, … Mouth) Driven Network Diffusion

Social
Networks
Stop Contagious Failures in Banking Systems

 During 2008 financial tsunami, which bank(s) we should inject

capital first to stop contagious failures in bank networks? 3
Utilize Peer Influence in Online Social Networks

 Influencer Marketing, Product Recommendation

 Who are the most influential people?
 What are the patterns of information diffusion?4
 Develop Strategies to Attack Terrorist Networks

A Global Salafi Jihad Terrorist Network

Hu et al. JHSEM 2009

 How to effectively break down a terrorist network? 5

 Business Networks Analytics and Applications

• Instructor: Prof. Dr. Daning Hu， BIN. 2.A.12

• TA: Xiao Li, BIN.2.A.24
• Email: hdaning@ifi.uzh.ch
• Credits: 5 ECTS credits
• Class Schedule:
http://www.ifi.uzh.ch/en/bi/teaching/Fall2017/NA.html
• Language: English
• Audience: Undergraduate and Master students
• Office Hours: Email for appointment, Room 2.A.12.

6
 Grading and Course Goals

• 1. One course project (90%)

• 2. Active participation and interaction during the lectures and

tutorials (10%)

• The project report should include the following four major

components:
– Network/Relational Data Collection (15%)
– Network Data Processing and Modeling (20%)
– Network Visualization (15%)
– Network Analysis (30%)

7
 Example 1: Network Data Collection

• Social Networks: Online communities, Social networking

websites, Personal blogs and micro-bloggings, online video
sharing websites. (e.g., Programmable Web)

• E-Business: Amazon Web Service, Ebay Data API, Taobao.

• Others: Financial, Education data sources: Stanford SNAP

Portal

8
 Example 2: Network Data Processing Modeling

• Extract relations/links from raw data in database tables

• Model such relations/links into network data.

– Node data
– Link data
9
 Example 3: Network Visualization
VMC MAR
DJ LTR HMT KSU
SO EOP EQR CSL
HMA
JNC EIX ROP
OXY
MWY MTD
DDIC KYO CPB ET CBSH SOTR GVA DST TLS UNH
T-O243 WTNYHMNGXPPOM SHB207
HIB FOX PKI CNA SPC
WMS
CAL
UB HCA EWBC
NWS BHI WRNC WCS NMGA PCGWAT KMX SKA213
ECL KEA KSE KBH BAN30
CNL
AVX ORCL
PXR DT USB TRNBSY EGN COM66
MHK ABX CDN
ICBC RTN NST GCI BBR
ATG
TW FJTSY ERICD
DOWFUWAY
TRZIGL TDW
DYN GMST DVN ASD
GWW AIB
EAGL INV116
KFY APCC BHP
FTS OEI.X CC NFG CLI
NBL EW
FIC MXT TSG VVI
STK TIN FBF MTB VFC WHR
ICN PLMD
HNCS.X AVY SNPS
RI SFG APD TGT PLD DV AZN
HUM MMS VVC
CNCEQ.OB EAS86 IMC NE MDP PL SMRT HCC DFS BSC
JAP120
NCI FLR CVS NTRS REG
GS LYG BBI
UCLRNA AMGN PKG AMCC CPY ONB TXU PEP
BKI STR AMAT BIR.X
IBE112TSN
ABT NT
JH ARG HPC BNI ITWPETM BP FRK CAH VIAB WFC AXPKWR
INTUAMZN JWN FL
COC.X ZION EAT
RY NAV BHMSQ.X PNX
TECUB BBV35 COP
LRCXCMX
WIND BCE HUG EDS S
AOC
WB GLT TFX
CWP AL OXM REGN LOW DRE SVU
CK RKT
CMA DTE EMR DEOMCD TRB GPT RMG
MAS
PCW BBDB MEH GPC BUD MAN AMR ALL BT
CLF N TOC EQT DELL BAC EXC PGN PSC
NEM WMO SUN
BRO JCI SLE KMB
TEF MWD AET MMM
BWS BID MODI MTX SBC SON NC SFA
HSIC ATK TDY
NSC PRU GE CVC
AEE GP CBE
RGA
HCR LECO SWY MI BLK WEC MAY PFE YUM MIT143
TD GIS KRI BSET HD
RSYS ENR BLL BDK BC TI STI X
AT CPN
ONE
BA BMY
RDK D GR PPE BLC
OLVTY AWE CPS
PHA.X MMC ATI CCC ROL MET
XOM WL GGC
AGL PGR BGG PNCEFX VZ JPM JCP
MRK HMC EC
SFAM.X LMT SEA203 WMB CYT
GKSRA OI SXT MIR MRO HON LNCE HUF DD
RAH CCE WYE UVN JHF
APH TXT TEL229 HNZ
BN TMO
MON HLT UAWGQ.OB
CHRW WFT AHC RPM
PNR NGG LXK SNRA KO COL PG MOLX
HUN.X
SUP TLAB GAP HSP
ENGSY PCO MDR
RUK JDEC GM MRBK
MIT144 TDS CB SEE DCN
CEG
TSSTEL233 GMH DAL PFGC LTXX OHP
ALN FON BLS
GRP
DLX
SNV PZB NOVL KRON FNFVRSN MKC
MSBHY WPO CUM
AA RX MGLH.OB CRS
BR NFX TP LVLT
CFWEQ.OB OSKWWW USM REY
RYN RSH POOL JNPR CAGBKS
GME
KR SFY AXL PRGX AV UCOMA ROOM
C CKR FPL
PCS KPN SYY WM AGRA EXPE TXCC TSAI CHEUY
LZ L AOL HIG CSX
CY TTEK T
TMCS.X CV VMI
RFMD SFD
IYCOY NDE USAI
NTAP FST PEG LU UTX CMCSA SPLS WBK B
DXYN
SAFCPOL VE MLHR ASTE HUWHY
UNP Q PPL CVX UNI249 OLN
SVELY COMS RGC SVM
CMOS
LLY TEN PTV DIS FDX ALD
BNK
V SCGLY UNM
CSCO
MHP GSK ALA RTP FTN ACF
IMI SAI192 ORB HET RIGS
OCR XRAY RRR BW WWY ROH AXASCH201
TOT
FNM XLNX ASO
AOT
TG FWC CBT MONI AVE FCX AMFI SUNW
CXR
ALB RRI LVMHY SZE TTN
GLW VRTY ADI ROK
PWN
TTI KEI IP PIN173 CQB PENX
NVO
NVLSCRUS TRW.X SJM CRE67 AGF TF1236 PX
ROY187 SI CYH
PLX DB DRSDY BNP41 EP
RTRSY REN177 BWG GSB.XASPX CLC
DAN73
CNP
KFT
NDSN
FD EON MUN146 FCEA FTI SOI AVA
TXI
COX BAN26 EAD85 DCX BF DUK
MDS PUK BAYZY SAN197 SSCC AFG
ALL14 AZ RWEOYBAY34 AIQUY ALK
SQAAACKHQ.OB
ERG93 FMC
ACLS HI.X ASH BMW40 AM RIGMGM
HGSI CRZBY SKS URI
MO URSMCK FLMIQ IR ION ABN VLKAY
IBM ESA AN
FISV MTG BK
CNF UL LORLY EMN
PCHRMK BCR AAUK NCC MGG
OMC
PAS MNY INET ROY188 DE AW
MDT GEN101 AKZOYVOD NSRGY
MWV NVS
CIN
KMRT RSG LGTO PCP
ETN PDLI CLX CRA CSR AICO
XRX SCH ASN DNZOYTIF
ABI SWI224 PH ATW BTI RIC180 SGP
AZPN GPS GT SHW
AMTD IT INGSNA
IVC
FITB DEL
APOL BEC AIG
TM
PBI CYTC
MAT ASPT
BOL FNVG.OB ELY UOPXVNO
ACTR.OB CHIR BSX AAPL SWK FOE
RAM.X HP BXS TRH
LUK SCS SWKS
BAX ACV PPDI MCDTA AMH ATH CNXT
LFG HRS CMS CTG SLB
CYGN OGE AGN
AIR HSY LNC FED
OCAS
FFBC ICOS BOKF
UVV
PLCM LII
HRH SRE DBD
R CSC

AVID

AVP

AVT

AVZ

AWK.X

AWR

AWRE

AXE

AXTI

AYE

AZO

AZR

BAESY

BAMM

BAN25

BAN29

BARZ

BBA

BBBY

BBOX

BBQCY

BBT

BBY

BCC

BCF

BCGI

BCM

BCP37

BCS

BDG

BDX

BE

BEAS

BEAV

BELFB

BELM

10
BEN

BER

BEV

BEZ

BFAM

BFB
 Example 4: Network Analytics

 A Microblogging Network: Who possesses the most

advantageous position in brokering information and knowledge in
this network? 11
 Example 5: A Global Terrorist Network

 How to effectively break down terrorist networks?

12
 Business Network Analytics or Applications

• Recommender Systems:

13
 Business Network Analytics or Applications

• Social Media based Marketing, Word-of-Mouth Effect

14
 Computing Tools Required In Tutorials

• Database Management Software

– MySQL or other common DBMS such as MS SQL Server, Oracle, etc.

• Network Visualization Tool: NetDraw, R (Statnet or iGraph)

• Network Analysis Tool: R (Statnet or iGraph), or UCINet

15
 Outline
 Syllabus

 Examples

 Introduction

 Social Network Analysis

 Social Network Data Modeling and Analysis

* Some of the contents are adapted from Prof. James Moddy’s slides at Duke University, and Prof
Jure Leskovec and Lada Adamic from Standford University 16
 Introduction: Why Study Networks?
 One of the most profound changes in today’s world is -
Decentralization
 Economical: BitCoin, Blockchain, P2P lending, etc.
 Social: Social Media News, Online Communities, Terrorist Cells, etc.
 Technological: Open Source Software, Virtual Teams, etc.

 The power, information, resources in real world networks are

becoming increasingly decentralized ->
 Nodes are distributed more equally, Less hierarchical;
 Good representations of entities and relationships in decentralized
systems.

17
The Focus: The Influence Mechanisms in Networks
 An average individual (node) can affect system outcome by
influencing its linked peers.

 This course will focus on the “influence” mechanisms that

network actors affect each other in terms of opinions,
behaviors, or risks through various ties, thereby changing the
network outcomes
 Economic influence: Risk contagion, etc.
 Social influence: Word-of-Mouth, observation learning, herding, etc.
 Network outcomes: Bank run, stock market crash, product diffusions.

18
 A “Random” History of Network Science

1736  Mathematical foundation – Graph Theory

1930
 Social Network Analysis and Theories
 Sociogram: Network visualization
 Six degree of separation
 Structural hole: Source of innovation
1990
 (Physicists) Complex Network Topologies
 Small-world model (e.g., WWW)
 Scale-free model (“Rich get richer”)
2000
 Network Science
 Economic networks (Agent modeling & simulation)
 Dynamic network analysis
 BI applications: product diffusion in social media,
2017 recommendation systems

 ? 19
 Network Science
• Network science is an interdisciplinary academic field which
studies complex networks such as information networks,
biological networks, cognitive and semantic networks, and
social networks. It draws on theories and methods including:
– Graph theory from mathematics, e.g., Small-world

– Statistical mechanics from physics, e.g., Rich get richer,

– Data mining and information visualization from computer science,

– Inferential modeling from statistics, e.g., Collaborative filtering

– Social structure from sociology, e.g., weak tie, structural holes

• Network science can be defined as "the study of network

representations of physical, biological, and social phenomena leading
to predictive models of these phenomena.”
20
A “Random” History: Math, Psychology, Sociology…
 The study of networks has emerged in diverse disciplines as
a means of analyzing complex relational data.

 Network science has its root in Graph Theory.

 Seven Bridges of Königsberg written by Leonhard Euler in 1736.
 Focusing on the properties of pairwise relations in a network
structure.

 Social Network Analysis

 Jacob Moreno, a psychologist, developed the Sociogram and to
“precisely describe the interpersonal structure of a group”.
 Stanley Milgram (Small World Experiment: Six Degrees of
Separation, 1960s). Facebook: 5.28 steps in 2008, 4.74 in 2011.
21
 “For the last thirty years, empirical social research has been
dominated by the sample survey. But as usually practiced, …, the
survey is a sociological meat grinder, tearing the individual from his
social context and guaranteeing that nobody in the study interacts
with anyone else in it.”
Allen Barton, 1968 (Quoted in Freeman 2004)

Moreover, the complexity of the relational world makes it

impossible to identify social connectivity using only our intuition.

Social Network Analysis (SNA) provides a set of tools to empirically

extend our theoretical intuition of the patterns that compose social
structure.
 The Origin of Modern Network Science:
Social Network Analysis
 Social network analysis (SNA) is a set of relational methods for
systematically understanding and identifying connections /ties
/relationships among actors.

 Social network analysis (SNA)

 is motivated by a structural intuition based on ties linking social actors

 is grounded in systematic empirical data

 draws heavily on graphic theory and imagery

 relies on the use of mathematical and computational models.

Jacob Moreno’s experiment on Friendship Network

Jacob’s experiment is the

first to use Social
Network Analysis
24
 What Does Social Network Analysis Study?

Social Network analysis lets us answer questions about social

interdependence. These include:

“Networks as Variables” approaches

• Are kids with smoking peers more likely to smoke themselves?
• Do unpopular kids get in more trouble than popular kids?
• Do central actors control resources?

“Networks as Structures” approaches

• What generates hierarchy in social relations?
• What network patterns spread diseases most quickly?
• How do role sets evolve out of consistent relational activity?

We don’t want to draw this line too sharply: emergent role positions can
affect individual outcomes in a ‘variable’ way, and variable approaches
constrain relational activity.
 Now…

Complex Networks in the Real World

Nodes Links

Social network People Friendship, kinship,

collaboration
Inter-organizational Companies Strategic alliance, buyer-seller
network relation, joint venture

Citation network Documents/authors Citations

Internet Routers/computers Wire, cable

WWW Web pages hyperlink

Biochemical network Genes/proteins Regulatory effect

… … …

26
 Now…

• Universal modeling and analysis methods for complex network

data

• Shared vocabulary between fields: Computer Science, Physics,

Sociology, Economics, Statistics, Biology

• “Big” Data availability: Internet, mobile, bio, health, security…

• Impact/usage: social networking, social media, marketing, etc.

27
 Our Approach

What Why How

 Social network  Econometric  Combine social

analysis (Metrics) identification of science methods,
casual Social and data mining, machine
 Describe the changes Economic influence learning with
in network evolution  Distinguishing econometric analysis
homophily
 Temporal changes in
network topological  Predict link formation
measures  Confounding
factors
 Simulate the
 Dynamic network evolution of networks
 PSM, DID, RD,
recovery
etc.

 (Relational) data  Explanations

mining 28
 What: Social Network Analysis

 Social network analysis (SNA) is a set of metrics and

methods for systematically describing, modelling, and
analyzing relationships among actors.

 Social network analysis (SNA)

 is motivated by a structural intuition based on ties linking
social actors

 is grounded in systematic empirical data

 draws heavily on graphic imagery

 relies on the use of mathematical and/or computational

models.
 What is a Network?

Actor/Node: Any entity in

a network
(person, system,
group, organization)

Tie/Link: Relationship
or interaction
between two nodes.
 Basic Concepts in (Social) Network Analysis

• Node, Actor, Vertex V

• Tie, Link, Edges E
• Network, System G (V, E)

• A link can be (1) Binary or Valued, (2)

Directed or Undirected.
b d b d

a c e a c e
Undirected, binary Directed, binary

b d b d
1 3 1 2
a c 4 e a c e
Undirected, Valued Directed, Valued
 Nodes or Social Actors

• Social Network data consists of two linked classes of data:

Nodes and Links.

• Node Example: Products in a purchase newtork

• Actor Examples: people in a group, departments within in a

corporation, public service agency in a city, nation-states in
the world system. “Node” does not imply that they have
intention or the ability to “act”.

• Network nodes are most often people, but can be any other
unit capable of being linked to another (schools, countries,
organizations, personalities, etc.)
 Links or Ties

• Actors (nodes) are linked to one another by social ties (links)

– Kinship, role-based, cognitive, affective, interactions, affliations

• Example of direct links in SNA (Wasserman/Faust 2008:17):

– Evaluation of one person by another (friendship, liking, or respect)
– Transfers of material resources (business transactions, lending or
borrowing things)
– Behavioral interaction (talking together, sending messages)
– Physical connection (a road, river, or bridge connecting two points)
Weight/Strength of Ties
Positive and Negative Weights
Two Modes of Social Network Analysis
One-mode Complete network

37
One-mode Ego network

38
 Ego Network Analysis

• Ego Network Analysis combine the perspective of network

analysis with the data of mainstream social science

• No computer assisted analysis needed

 Two-mode Complete Network (Bipartite Graph)

40
 Two-mode (Bipartite) Network Transformation

 From Zan Huang et al., 2009, Management Science

41
 Network Data Modeling: Adjacency Matrix

42
 Network Distance (Weighted) Adjacency Matrix

43
 Major Network Data Formats (in UCINet)

44
Real World Networks are Sparse Graphs
More Types of Networks