Introduction to Network Analysis saverio . giallorenzo @gmail.com 1 - - PowerPoint PPT Presentation

▶
introduction to network analysis
SMART_READER_LITE
LIVE PREVIEW

Introduction to Network Analysis saverio . giallorenzo @gmail.com 1 - - PowerPoint PPT Presentation

Feb 04, 2023 854 likes •1.27k views

Web Science Introduction to Network Analysis MA Digital Humanities and Digital Knowledge, UniBo Introduction to Network Analysis saverio . giallorenzo @gmail.com 1 Web Science Introduction to Network Analysis MA Digital Humanities and

slide-1
SLIDE 1

saverio.giallorenzo@gmail.com Web Science • Introduction to Network Analysis MA Digital Humanities and Digital Knowledge, UniBo

Introduction to Network Analysis

1

slide-2
SLIDE 2

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

2

slide-3
SLIDE 3

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 3

Networks

slide-4
SLIDE 4

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

4

They love each other Juliet’s father Romeo’s father They hate each other Romeo’s best friend They planned a ruse Juliet’s cousin Juliet best friend

slide-5
SLIDE 5

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 5

Networks

A network is a simplified representation that reduces a system to an abstract structure

  • r topology, capturing only the basics of the

connection patterns and little else.

slide-6
SLIDE 6

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 6

Networks

  • Networks capture the pattern of interactions between the parts of

a system. In turn, the pattern of interactions have a sensible effect on the behaviour of a system.

  • Examples:
  • the pattern of connections between computers on the Internet

affects the routes that data take over the network and hence the efficiency with which the network transports those data.

  • the connections in a friendship network affect how people

learn, form opinions, and gather news, as well as other less

  • bvious phenomena, such as the spread of disease.
  • Knowing the structure of a network is essential to fully

understand how its corresponding system works.

slide-7
SLIDE 7

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 7

Networks, examples

slide-8
SLIDE 8

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 8

Networks

The systems studied can have interesting features not represented by the network—e.g., the detailed behaviours of individual nodes, such as people and the precise nature of the interactions between them.

slide-9
SLIDE 9

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 9

Networks, examples

Mercutio Paris Prince Benvolio Friar Laurence The Nurse Lady Capulet Lord Capulet

Lord Montague Lady Montague

Apothecary

Friar John

Romeo Juliet

Tybalt

slide-10
SLIDE 10

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 10

Networks

We can capture additional information by labelling the nodes and/or edges of the network, such as with names or strengths of interactions.

slide-11
SLIDE 11

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Mercutio Paris Prince Benvolio Friar Laurence The Nurse Lady Capulet Lord Capulet

Lord Montague Lady Montague

Apothecary

Friar John

Romeo Juliet

Tybalt

11

Networks, examples

kills cousin friendship in cahoots servant/friend

  • ffspring

married married

  • ffspring

dealership friendship in love in feud lost

slide-12
SLIDE 12

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 12

Networks

Finding what is the “right” kind/amount of information to make a system treatable (to reasoning) is a work of craftsmanship and experience. The invariant here is that, every time we define a representation

  • f a full system, we decide to filter out some information.
slide-13
SLIDE 13

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Mercutio Paris Prince Benvolio Friar Laurence The Nurse Lady Capulet Lord Capulet

Lord Montague Lady Montague

Apothecary

Friar John

Romeo Juliet

Tybalt

13

Networks, examples

♠ ❤ ❤ ❤ ❤ ❤ ♣ ♣ ❤ ❤ ❤ ❤ ❤ ❤ ♠

slide-14
SLIDE 14

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Juliet

14

Networks, examples

kills cousin friendship in cahoots servant/friend

  • ffspring

married married

  • ffspring

dealership friendship in love in feud lost

slide-15
SLIDE 15

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 15

Networks, examples

slide-16
SLIDE 16

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

16

slide-17
SLIDE 17

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

17

slide-18
SLIDE 18

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

18

slide-19
SLIDE 19

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

19

slide-20
SLIDE 20

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

20

slide-21
SLIDE 21

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

21

The Internet map

slide-22
SLIDE 22

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

22

Cina U.S.A Russia Japan Italy Ukraine France Germany United Kingdom Brazil Poland Iran Spain India

slide-23
SLIDE 23

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 23

Network Analysis

The founding questions behind network analysis are:

  • If we know the shape of a network, what can we learn about

the nature and function of the system it describes?

  • How are the structural features of a network related to the

practical issues we care about?

slide-24
SLIDE 24

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 24

Network Analysis

A first step in analysing the structure of a network is often to make a picture of it. Automatic tools help in managing, visualising, and exploring networks. Visualisation is a useful tool in the analysis of network data, allowing us to instantly see important structural features that would otherwise be difficult to pick out of the raw data. The human eye is enormously gifted at discerning patterns, and visualisations allow us to put this gift to work on our network problems.

slide-25
SLIDE 25

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 25

Networks, examples

slide-26
SLIDE 26

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 26

Network Analysis

While we are good at spotting patterns, we can feasibly do that manually up to a few hundreds or thousands of nodes and for networks that are relatively sparse—whose number of edges is quite small. To address these issues, network theory has developed a large tool-chest of measures and metrics that “mimic” some specific abilities of our eyes, to help us when visualisation is impossible or unreliable.

slide-27
SLIDE 27

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 27

Network Analysis

  • An example of a useful (and widely used) class of network

measures are the centrality measures.

  • Centrality quantifies how important nodes are in a network.
  • While the concept is clear, what it (mathematically) means for a

node to be central in a network may vary.

  • The simplest centrality measure is called degree. The degree of

a node in a network is the number of edges attached to it.

  • In many cases, the nodes with the highest degrees in a network

also play major roles in the functioning of the system. Hence the node’s degree can be a useful guide to focus our attention on a system’s most important elements.

slide-28
SLIDE 28

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis

Networks, examples

28

Google Twitter Facebook Microsoft Yahoo Ask.com

slide-29
SLIDE 29

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 29

Network Analysis

  • Another example of a network concept that arises repeatedly

and has real practical implications is the so-called small-world effect.

  • Given a network, one can ask what the shortest distance is,

through the network, between a given pair of nodes. In other words, what is the minimum number of edges one would have to traverse in order to get from one node to the other?

  • Although first studied in the context of friendship networks (the

famous “six degrees of separation”), small-world effect appear to be widespread, occurring in essentially all types of networks.

slide-30
SLIDE 30

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 30

Networks, examples

slide-31
SLIDE 31

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 31

Network Analysis

An example of small-world configuration is the occurrence of clusters or communities of a small number of individuals linking the majority of nodes in the network

slide-32
SLIDE 32

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 32

Networks, examples

slide-33
SLIDE 33

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 33

Networks, examples

slide-34
SLIDE 34

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 34

Networks, examples

slide-35
SLIDE 35

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 35

Network analysis

  • Scientists in a wide variety of fields have developed an

extensive set of mathematical and computational tools for analysing, modelling, and understanding the current status of a network (e.g., which is the best connected node or how similar two nodes are to one another) and make predictions about processes taking place on networks (e.g., the way a disease will spread through a community).

  • Since networks abstract specific details of the system they

represent, the same mathematical tools can be applied to almost any system that has a network representation (the power

  • f abstraction).
slide-36
SLIDE 36

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 36

This course (1/2)

  • We study applications of network analysis on diverse fields—

literature, philosophy, biology, computer science, economics, and forensics (but not only limited to those)—to understand both the motivations and techniques employed;

  • We acquire proficiency in information literacy to learn how to

discover network information and develop an understanding of how that information is produced and its use in creating new knowledge;

  • We study relevant theory of network analysis;
slide-37
SLIDE 37

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 37

This course (1/2)

  • We learn how to perform the research design of a network

analysis study, divided in its primary elements:

  • formulation of theories and definition of hypotheses;
  • data collection and data management;
  • identification and application of network measures;
  • visualisation and interpretation of the measure results.
slide-38
SLIDE 38

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 38

This course (2/2)

  • This is a fairly interactive course:
  • we will see the application of the theory with automatic tools;
  • students will form groups, analyse research papers, and give

seminars to the rest of the class.

slide-39
SLIDE 39

saverio.giallorenzo@gmail.com MA Digital Humanities and Digital Knowledge, UniBo Web Science • Introduction to Network Analysis 39

The exam

  • Possible starting points are:
  • exploring and interpreting the

characteristics of some network. E.g., “to understand the dynamics of network X, I apply measures Y, W, and Z, and give an interpretation of the results, following some related studies in the literature.”

  • finding predictors. “in the context of

X is relation Y a predictor for phenomenon Z?”

  • proposing new measures: “I present

measure X which is an indicator of Y in a network of shape Z.”

  • The exam consists in a report
  • n a network analysis research

designed and performed by the student, individually.

  • The subject of the research, the

goals, and the data are agreed upon with the teacher near the end of the course.

  • Start thinking of a couple of

possible ideas (so you do not rush it toward the end of the course).