Social Network Analysis in Intelligence

1
Social Network Analysis in Intelligence

• Emily Keefe
• Policy and Capability Studies

2
Presentation overview

• Introduction to SNA
• SNA in Government and Dstl
• Intelligence SNA

3
Social Network Analysis (SNA)
4
The history of SNA
Psychology
Anthropology
Mathematics
5
Genres of Social Network Analysis
?(cmax - c(vi))
Complexity
Graph theory
Structures
Tools
6
?(cmax - c(vi))
Structures
Tools
7
Origins of SNA applications in Dstl
Modelling target dynamics
Modelling the critical national infrastructure
Bayesian inference for link prediction
Social networks of terrorist groups
Proof of concept study
Social Network Analysis Case Study
Cross-Government Case Study
8
SNA in Government
X
?(cmax - c(vi))
9
Why use SNA?

• Helps people understand the groups they are
  interested in.
• SNA may identify network characteristics not
  apparent by eye
• SNA allows group structures and flows of
  communications to be explored
• Systematic approach to data coding allows for
  quick and easy what-if? analysis
• Coding data forces people to challenge their
  assumptions and allows knowledge to be transferred

10
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
11
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
12
Set up

• Getting a tasking
• Identifying the team
• Determining aims questions to be answered
• Deciding on a taxonomy
• Sharing out responsibility
• Selecting and obtaining software
• Creating a suitable database

13
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
14
Coding

• Finding people to code
• Holding an initial training session or workshop
• Regular team meetings
• Coder interviews
• Inter-coder reliability exercises

15
Why coding and definitions matter

• Is A co-located with B if
• They work in the same building?
• They work on the same site?
• They work for the same company?
• They attend the same meeting?
• They meet and talk on the street?
• They were in jail together?

16
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
17
Data management

• Configuration control
• Database merging
• Database refinements
• The human element
• Data tidying
• Information entered into the wrong field
• Locations not entered
• Dates associated with data entered
• Entity disambiguation

18
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
19
Questions and Techniques
20
Intelligence SNA
Data management
Present results
Set up
Coding
Analysis
21
Why use SNA?

• Systematic approach to data coding allows for
  quick and easy what-if? analysis
• Coding data forces people to challenge their
  assumptions and allows knowledge to be
  transferred
• SNA may identify network characteristics not
  apparent by eye
• SNA allows group structures and flows of
  communications to be explored

22
Centrality measures

• Used to explore who is key in the network

Eigenvector Centrality Who does who you know
know?
Degree Centrality Well connected hubs
Betweenness Centrality - Gatekeepers
23
What about missing data (1)?
24
What about missing data (2)?
25
What about deliberately concealed data?

• Is there anything we can do?
• Should we compensate for missing and concealed
  data, or simply accept that it exists and try to
  minimise the impact it has?

26
Reachability and distance
27
Affiliations inferring links through shared
attributes
Groups
People
28
Reachability and distance
29
More trustworthy links
30
Reachability and distance
Normal link
Trusted link
31
Difficulties encountered
Software packages available
Balance between coding and analysis
Nature of available data
Applicability of SNA measures
Continual process of data tidying
32
Demonstration of added value

• Identification of a key individual through
  inferred links
• Identification of a different inter-group
  structure to that which had been assumed
• Understanding of the importance of an individual
  of interest

33
Assumed structure
Group A
Group B
34
Structure suggested by SNA
Group A
Group B
35
Links of a person of interest
36
(No Transcript)
37
The history of SNA
Psychology
Anthropology
Mathematics
38
Review of Terrorist disruption
Historical CT success factors
Exploitation Group analysis and assessment /
DIS, PJHQ HQ ARRC Operational assessments
AB Analysis
AB Assessment Metrics
SNA Case studies
Influence Analysis / PAM
Media assessment
Media monitoring
Suicide Terrorism
Computer-assisted sentiment analysis
Modelling Info Ops Target Dynamics
Leadership assessment
Coercion / Intent gaming
Prediction of surprise attack
Body movement profiling
LAPA RV Assessment
Exploitation Remote personality assessment
DIS Leadership and Personality Assessment (LAPA)
39
Proposed ISNA Activities
Support development of Behavioural, SNA andSNEE
Science Base
Technical / softwaresupport
ISNACapabilityDevelopment
Operational ExploitationDstl
Operational ExploitationMoD, OGAsand OGDs
End 2007
Current
End 2010
40
QuestionsComments