Online social networks and social media play a central and growing role in our daily life. They influence how we communicate with our friends, how we obtain and consume media and news content, how we apply for a job, even how the information we look up in Wikipedia is generated or how political parties try to convince you to vote them.

The aim of this course is to provide both the theoretical background as well as the practical tools to analyse, model and visualize the multiple facets of these phenomena.

We will give an overview about the state of the art in Social Network Analysis and how its metrics have been derived from theories in Sociology. Furthermore we will provide knowledge about tools and methods to derive specific social media datasets and the way to visualize social networks that go beyond an ugly hair ball of nodes and edges. Finally we will also treat more advanced topics such as learning and inference, information diffusion, community structure and prediction.


The course will have a strong empirical component, mixing seminar lectures with hands-on analysis of real-world data sets. The student will incrementally develop a project based on a real-world dataset to be presented at the last session of the course. The evaluation will be based on weekly exercises (50%) and a final project presentation (50%).


Session 1 (Vicenç / Andreas): Overview and Introduction. Social Distance (Small World, Average/Maximum path lengths). Ego Networks (degree distributions, Dunbar's number, clustering coefficient). Social Influence (homophily, assortativity, centrality measures). Weak ties.

Session 2 (Vicenç): Network formation models I: random network models, preferential attachment, hybrid models, triadic closure, parameter estimation.

Session 3 (Andreas): Data acquisition: Hands-on working on real datasets, Available tools and libraries for social network analysis. SNAP, igraph, nodeXL and neo4j. Dataset repositories. 

Session 4 (Vicenç): Network formation models II: Strategic formation models, choices, utility functions, stability vs efficiency, connections and co-author models, parameter estimation.

Session 5 (Andreas): Social Network Visualization, Available tools, Pitfalls, Introduction to Gephi. Choice of dataset for project.

Session 6 (Vicenç): Dynamics and Behaviour: diffusion, adoption curves, epidemics: SIR/SIS models, social influence. Fake news and social media, sources of bias, bot detection, examples.

Session 7 (Andreas): Community discovery: Graph partitioning, hierarchical clustering, divisive algorithms. modularity-based methods. Evaluation measures. Applications.

Session 8 (Vicenç): Statistical modeling of online discussions. Existing models and applications: platform design, prediction of user behavior, role of content, influencing user activity.

Session 9 (ChaTo): Natural experiments - randomized controlled experiments, Neyman's model; Matching studies - matching and propensity matching design.

Session 10 (Students): Project presentations.



[1] Networks, Crowds, and Markets: Reasoning about a Highly Connected World, David Easley & Jon Kleinberg. Cambridge University Press (2010).

[2] Social and Economic Networks, Matthew O. Jackson. Princeton University Press (2010).

[3] Social Network Analysis for Startups: Finding connections on the social web. Maksim Tsvetovat, Alexander Kouznetsov. O'Reilly Media (2011).