This module explores advanced graph types, including bipartite, weighted, temporal, and scale-free networks common in social media platforms. Students implement fundamental graph algorithms like DFS, BFS, and Dijkstra's algorithm for network exploration and shortest path analysis. The module covers network connectivity, components, and global properties such as density and efficiency. Students learn to analyse network structures and understand algorithmic complexity considerations for large-scale social media networks. Practical demonstrations guide students through implementing graph algorithms and analysing real social media network properties using computational tools.

This module focuses on measuring node importance and identifying influential users in social networks. Students master fundamental centrality measures including degree, betweenness, closeness, and PageRank algorithms to analyse user roles and network positions. The module covers local node properties, structural patterns like transitivity and homophily, and link prediction techniques. Students learn to profile users based on multiple network measures and understand social network formation principles. Hands-on demonstrations teach students to compute centrality measures and build comprehensive user analysis systems for social media applications.

This module examines methods for identifying and analysing groups within social networks. Students explore community detection approaches, including modularity-based methods, the Louvain algorithm, and spectral clustering techniques. The module covers overlapping communities, dynamic community evolution, and quality evaluation metrics. Students learn to compare different detection algorithms and understand their strengths and limitations. Applications in targeted marketing, content recommendation, and information flow analysis are emphasised. Practical demonstrations guide students through the implementation of community detection algorithms and the analysis of community structure in real social media networks.

This module studies how information and behaviours spread through social media networks. Students explore diffusion models, including independent cascade and linear threshold mechanisms, along with influence maximisation techniques. The module covers collective behaviours such as herd mentality, echo chambers, and social contagion phenomena. Students learn to detect information cascades, distinguish influence from homophily, and predict viral content. Applications in crisis detection, marketing campaigns, and behaviour prediction are emphasised. Comprehensive demonstrations teach students to simulate diffusion models and analyse real-world information spread patterns.

This module describes the design of recommender systems in the modelling of social media. The application domains for the recommendation models and systems are summarised. The Internet-scale algorithms using rule-based and parameter-based techniques are given. Further optimisation based on recent advancements in deep learning is also discussed. The various data analytics tasks in the recommendation problems are given based on the previously studied data mining models, such as clustering, frequent pattern mining, and association rule mining.

This module provides the characterisation of big data generated on social media platforms. It provides an introduction to the adaptations of the data analytics tasks for processing big data. Complex graph analysis is explained in terms of dynamic networks formed on social media datasets. The corresponding mathematical properties to be satisfied by the complex datasets, such as non-stationarity and causality, are then incorporated into the data analytics algorithms. The resultant downstream applications are discussed with reference to recent developments in Agentic AI. Emerging technologies based on AI robustness and fairness are also introduced with reference to misinformation, disinformation, and the weaponisation of social media in multi-stage cyber attack campaigns.

This module introduces robust and privacy-aware algorithm design for social media systems operating under adversarial conditions. It covers polarization mitigation through network interventions and adversarial perturbations, misinformation detection, encryption and anonymization techniques, reinforcement and bandit learning for adaptive recommendation, and hybrid deep learning models. The module also integrates MLOps practices for deploying, monitoring, and maintaining responsible ML-driven platforms.

This module examines advanced content-based and personalised recommendation frameworks in social media ecosystems. It introduces language modelling, topic modelling, and novelty-detection filters for analysing multilingual and multimedia content. Knowledge representation through semantic web architectures, ontology engineering, and knowledge graphs is integrated with web data mining techniques. The module further explores transformer architectures, foundation models, multimodal learning systems, and adversarial deep learning in black-box environments. Privacy-preserving analytics and adversarial attacks on data privacy models are discussed within the broader context of responsible and scalable social media intelligence systems.

This module examines algorithmic approaches to detecting and countering malicious content and adversarial behaviour in digital ecosystems. It covers fake news characterization, misinformation propagation patterns, feature engineering, and graph-based detection models. Image forensics and deepfake generation and detection are analysed using adversarial learning, geometric features, and decision boundary sensitivity analysis. The module further explores cyberbullying detection, phishing and URL analysis, information warfare strategies, and adversarial deep learning in attack–defense scenarios. Applications include LLMs, foundation models, wargaming simulations, multi-agent systems, and game-theoretic frameworks for AI-driven cybersecurity and strategic decision-making environments.