Instructors: Dr Riccardo De Vita (University of Greenwich) and Yasaman Sarabi (University of Greenwich)

This course is aimed at those researchers and post-graduate students who are new to the field of social network analysis (SNA), and would like to better understand whether and how they can use it to enhance their research programmes. The goal of the course is to provide attendees with a general overview of the field of social network analysis, confidence in using its key analytical tools in practice, and insight into how it can be used in scholarly practice in different fields. Focus is on research design and how SNA elements can be successfully integrated into a research project, paper, or dissertation.

**Requirements**

All social science backgrounds are welcome, and participants are assumed not to have any previous knowledge of SNA, or of any analytical or statistical software. Participants will be mainly expected to use UCINET and Netdraw during the course (even if other software will be discussed). No previous experience with the software is expected.

**Learning outcomes**

At the end of the course participants will be able to: (1) independently design a research programme requiring SNA in their own field of research; (2) ollect and manage network data; (3) analyse, interpret and visualise fundamental network measures at the individual, group and network level; (4) confidently use UCINET and NetDraw to perform network analysis and visualise network data.

The goal of the course is to provide attendees with a general overview of the field of social network analysis, confidence in using its key analytical tools in practice, and insight into how it can be used in scholarly practice in the social, economic, managerial and political disciplines. Focus is on research design and how SNA elements can be successfully integrated into a research project, paper, or dissertation.

**Indicative content**

The content below is indicative. The course instructor will welcome input and requests from class participants. Time will be allowed for discussion about individual research projects and interests. Individual consultation outside classroom activities can also be scheduled.

Day 1 – Introduction to SNA

Introduction to Network Research

Network Terminology and Data

Introduction to UCINET and NetDraw.

Day 2 – Designing Network Research

Data collection, management and ethics

Two-Mode Networks

Data Management in Ucinet and NetDraw

Day 3 – Individual Level of Analysis: Centrality and Brokerage

Centrality measures

Structural Holes

Applications in Ucinet and NetDraw

Day 4 – Group and Network Level of analysis

Identification of subgroup techniques

Characterising Networks

Applications in Ucinet and NetDraw

Day 5 – Special Topics

To be agreed with class participants

Instructor: Dr Guido Conaldi (University of Greenwich)

Social networks are dynamic by nature. For example, network dynamics are important for domains ranging from friendship networks (e.g., van Duijn et al., 2003; Burk et al., 2007) to inter-organisational networks (Borgatti and Foster, 2003; Berardo and Scholz, 2010; Agneessens and Wittek, 2011). Ties can be established and can be terminated; also there may be changes in the actors taking part in the network. Changes in ties may be considered the result of the structural positions of the actors within the network - e.g., when friends of friends become friends -, characteristics of the actors ('actor covariates'), characteristics of pairs of actors ('dyadic covariates'), and residual random influences representing unexplained influences. The study of network dynamics sheds light on the underlying theoretical micro-mechanisms that induce the evolution of social network structures on the macro-level.

**Requirements**

While no prior experience with dynamic models for social networks is assumed, some familiar with basic social network concepts and terminology will be useful. This course is aimed at researchers and students willing to explore the dynamics of social networks.

**Learning outcomes**

Social scientists having the possibility to access and exploit the wealth of longitudinal datasets on social networks that are becoming available require analytical tools able to cope with the scale and complexity of such data. To this purpose the course will introduce participants to the techniques and tools required to visualise and analyse (1) panel network data using the Stochastic Actor-Oriented Models - a.k.a. SIENA Models - (Snijders et al., 2010) and (2) event network data using the Relational Events Models (Butts, 2008). The course will also introduce and discuss the main methodological and empirical studies that in various fields of social sciences have contributed to and successfully applied such methods.

Instructor: Dr. Nicola Perra (University of Greenwich)

The unprecedented amount of data now available in many disciplines changed completely the way we look, understand and study the world and its properties. Network Science provides the new paradigm to analyse and deal with this data deluge.

This course will introduce students to mining and analysis techniques in Network Science. The students will learn about working with real world datasets and their description as networks. More precisely, the course will provide an overview on Network and Data Science. The most important properties and metrics will be discussed.

**Requirements**

There are no prior requirements for this workshop.**Learning outcomes**

This course will provide a basic introduction to python that will be used throughout the module. The students will be guided through two “data-hands on” case studies. The first will be centred on a real dataset describing face-to-face interactions collected with RFID tags in the project "Sociopatterns". In this example, by using python the students will learn how to transform datasets into network data structures and how to compute their metrics and measures. The second instead, will revolve around a dataset of geo-localized tweets. By using python the students will learn how to build different types of networks considering as nodes topics or countries.

Furthermore, the students will learn basic elements of clustering and community detection techniques that will be directly applied to the different Twitter networks. The course will also discuss the basic ingredients of data mining necessary to gather data from Twitter and other online platforms. In the final part of the module students will then learn how to model dynamical processes unfolding on networks such as the spreading of infectious diseases on face-to-face interaction networks or the diffusion of memes on Twitter.

Instructors: Dr. David Dekker (University of Greenwich) and Dr Francesca Pallotti (University of Greenwich)

This course covers Quadratic Assignment Procedure-based (QAP-based) approaches and Exponential Random Graph Methods (ERGMs) for analysis of cross-sectional network data.

Central in this course are applications of various QAP-based approaches that try to deal with issues concerning the issues inherent to network data. The usefulness of QAP-based approaches is that they aim to leave intact the estimation method, but replace inference based on classic statistical tests with inference based on randomization tests. First, an insight is created in the underlying issues that prevent classic statistical tools to be applied in case of network data. Subsequently, a treatment is given on how different statistical methods of multivariate-analyses can be adjusted so that they can be applied to network data. The course hence builds on and extends the knowledge of researchers, so that they can apply much of their existing toolkit to network data.

The course also introduces participants to the general theoretical background of ERGMs. These models are uniquely valuable in their ability to connect local structural configurations induced by specific social mechanisms that are not directly observable (e.g., reciprocity, transitivity) to global features of observed network data (e.g., clustering, community structures). The technical aspects of model specification, estimation, goodness of fit, and parameter interpretation are then tackled with examples of empirical applications and hands-on exercises.

The software packages that used in the course are UCINET 6, R, MPNet.

**Requirements**

While no prior experience with QAP-based approaches and ERGMs is assumed, participants are expected to be familiar with social network concepts and terminology, and to have some knowledge of basic concepts in statistical inference. The main reference book for the workshop is: Lusher D., Koskinen, J., and Robins, G. (Eds.) 2013. Exponential Random Graph Models for Social Networks: Theories, Models and Applications. Cambridge University Press, New York. For an introduction to ERGMs participants may want to read: Robins, G., Pattison, P., Kalish, Y., and Lusher, D. 2007. An introduction to exponential random graph (p*) models for social networks. Social Networks 29: 173–191.

**Learning outcomes**

At the end of the workshop participants will: (1) be aware of the potential benefits offered by QAP-based approaches and ERGMs for the analysis of social networks between individuals and between organizations; (2) be able to develop research hypotheses that can be tested by using QAP-based approaches and ERGMs; (3) be able to specify, estimate and interpret QAP-based model ERGMs.

**Indicative content**

Day 1 – Introduction to Hypothesis Testing in SNA

Introduction to Statistical Analysis of Social Network Data

Why Do Inferential Statistics?

Dealing with Data Dependencies: Permutation Tests

Applications in UCINET and R

Day 2 – Linear Models in SNA

Multivariate Regression Analyses: A Reminder

Permutation Testing for Multivariate Models

Multicollinearity and Skewed Data

The Linear Probability Model and Heteroskedasticity

Applications in UCINET and R

Day 3 – Permutation Schemes, Time Series Data, and Network Effects

Known Clustering and Blocked Permutations

Longitudinal Models

Nodal, Dyadic and Higher-Order Nodal Effects

Applications in UCINET and R

Day 4 – Exponential Random Graph Models

ERGMs: Rationale

Working with graph distributions: ERGMs and dependence

General form of ERGMs

Applications in MPNet

Day 5 – Applying ERGMS in Practice

Estimation, Interpretation and Goodness of fit

Recent extensions

Testing Trust: An Example

**2-day short course on 13-14 June**

Instructor: Dr. Rossano Schifanella (University of Turin)

Machine learning has become an integral part of our everyday life changing the way humanity addresses an endless set of problems in many disciplines, bringing us - for example - self-driving cars, speech recognition systems, effective web search, algorithms that outperform humans in several tasks, and a vastly improved understanding of the human genome.

This course will provide an understanding of the basic principles of machine learning and it will guide the students in developing and evaluating practical solutions in a real world setting.

**Requirements**

There are no strict prior requirements for this course. Basics of programming in Python is suggested to follow the practical sessions, introductory knowledge on probability theory and linear algebra are a plus.

**Learning outcomes**

In this introductory course, you will learn about some of the most effective machine learning techniques, and gain practice to apply them in real case scenarios. The students will be guided through the theoretical underpinnings of learning and will acquire the practical know-how needed to solve real world problems using Python and open source libraries. Hands-on sessions will provide a step-by-step overview of the whole learning process, providing a practical overview on how to manipulate real dataset, model a learning problem, find the most appropriate solution, and evaluate the results. At the end of the course, students will be able to deal with supervised learning approaches (classification, regression models using parametric/non-parametric algorithms, support vector (machines), unsupervised learning (clustering, dimensionality reduction) and they will gain basic knowledge of the best practices and evaluation metrics used in the field.

Machine learning has become an integral part of our everyday life changing the way humanity addresses an endless set of problems in many disciplines, bringing us - for example - self-driving cars, speech recognition systems, effective web search, algorithms that outperform humans in several tasks, and a vastly improved understanding of the human genome. This course will provide an understanding of the basic principles of machine learning and it will guide the students in developing and evaluating practical solutions in a real world setting.