Network Analysis (Autumn 2019)

Abstract

The course is focused on forming the basis of a network scientist. The final objective is to have the students being able to fully appreciate the difficulties of the problem of finding communities in social networks. To achieve this objective, a complete knowledge of network science is required. Each concept necessary to understand communities in networks has to be fleshed out in previous lectures, and the concepts on which it depends have to be presented beforehand.

Description

Networks science is a thriving field of study. The reason of its popularity is its ability to represent very complex phenomena with the very simple model of a graph. With network science one can analyze global societal patterns, catastrophic breakdowns of distributed infrastructures, the metabolic pathways in humans at the basis of diseases, among many examples.

1. Challenges & applications of network science: the many surprising properties of real-world networks.
2. Linear algebra applications to network analysis.
3. More graph properties & fitting power law degree distributions.
4. Graph models.
5. Spreading processes.
6. Link prediction.
7. Exam simulation #1.
8. How to deal with noisy data: projection, backboning, and sampling.
9. The mesoscale: homophily, core-periphery, and hierarchies.
10. Graph mining.
11. Community discovery #1: partitions & evaluations.
12. Community discovery #2: overlaps & hierarchies.
13. Community discovery #3: bipartite & multilayer.
14. Exam simulation #2.

Formal prerequisites

The course is required for BSc students of Data Science third semester. 
The student must always meet the admission requirements of the IT University.

Intended learning outcomes

After the course, the student should be able to:

• Define various types of network structure and calculate the main descriptive metrics
• Describe the main characteristics of a given network structure
• Analyse real world social networks and summarize their main characteristics and how those affect network-based processes (e.g. propagation)
• Identify community structures in network and provide an appropriate interpretation of these structures.

Learning activities

14 weeks of teaching consisting of lectures and exercises

1. Challenges & applications of network science: the many surprising properties of real-world networks.
2. Linear algebra applications to network analysis.
3. More graph properties & fitting power law degree distributions.
4. Graph models.
5. Spreading processes.
6. Link prediction.
7. Exam simulation #1.
   
8. How to deal with noisy data: projection, backboning, and sampling.
9. The mesoscale: homophily, core-periphery, and hierarchies.
10. Graph mining.
11. Community discovery #1: partitions & evaluations.
12. Community discovery #2: overlaps & hierarchies.
13. Community discovery #3: bipartite & multilayer.
14. Exam simulation #2.
    

Mandatory activities

Groupwork project. Successful completion of the group project is a requirement for admission to the exam.

The student will receive the grade NA (not approved) at the ordinary exam, if the mandatory activities are not approved and the student will use an exam attempt.

Course literature

“Network Science” by A-L. Barabasi (free book available at http://networksciencebook.com/). 

Additional material (teacher's notes) provided via LearnIT.

Ordinary exam

Exam type:
A: Written exam on premises, external (7-trinsskala)
Exam variation:
A22: Written exam on premises with restrictions.
Exam description:

Duration: 4 hours. 

Restrictions: Only NetworkX, Graph Tool, and/or iGraph and an IDE can be used during the exam.

Books and notes cannot be used during the exam.

reexam

Exam type:
Z. To be decided, - (-)

Time and date