How does information spread? How can we predict social interactions? Addressing these questions requires analyzing massive volumes of data. Many complex datasets can be naturally represented as graphs that capture relationships and interactions among entities. Graph-structured data can then be processed by advanced algorithms to perform tasks including classification, clustering, and forecasting. This course examines the computational, algorithmic, and modeling challenges involved in analyzing large-scale social graphs. By exploring the structural properties of those graphs, you will learn how to apply machine learning and data mining techniques to uncover patterns, improve predictions, and gain insights across a wide range of networked systems.
The course begins with an overview of the fundamental concepts in social media and network analytics, including graph-based data representations, similarity and centrality measures, network metrics, and graph storage and processing tools. Building on this foundation, the course explores a range of algorithms for analyzing network structure, with particular emphasis on community detection, as well as methods for predicting missing or future links and nodes in social networks. Students will then study additional important applications of social network analysis, such as information diffusion and influence modeling, social recommendation systems, social and network mining, and knowledge graphs. Throughout the course, both theoretical principles and practical applications will be emphasized to provide students with a solid understanding of how social network analytics is applied in real-world scenarios.
The course is structured to include both a deep theoretical component and a hands-on practical component. The theoretical part focuses on the fundamental principles, models, and algorithms underlying social network analysis and social media analytics. Complementing this, the practical part is delivered through a semester-long project in which students apply the concepts and techniques learned in class to real-world social network data. Through this project, students gain experience in data collection, analysis, implementation, and evaluation, bridging theory and practice and developing practical problem-solving skills.
This course consists of lectures and hands-on exercises/labs. The weekly/bi-weekly exercises are an important part of the learning process. Solving the exercises will be the best way to prepare for the final exam.
The lectures take place TU 14:15-17:00 in room E130 (UniFR, PER17). The lecture notes for the course will become available as we progress through the semester.
The textbook for the course is Social Data Mining: an Introduction , First edition, Cambridge University Press, Reza Zafarani, Mohammad Ali Abbasi, and Huan Liu, 2014.
As part of the course assessment, the students need to deliver a semester-long project. The general aim is to analyze a large social media dataset and uncover its underlying characteristics. Students must develop a tool that integrates the entire social data pipeline, from data collection through analytical visualization. The project should include implementing social graph analytics from scratch and evaluating their performance in comparison with existing solutions (whether from the course, research literature, or other sources).