Adrien Guille's PhD thesis focuses on understanding and analyzing information diffusion in online social networks. The key contributions of this work include: 1. **Survey of Developments**: A comprehensive review of existing methods and models for information diffusion, including popular topic detection and influential spreader identification. 2. **T-BaSIC (Time-Based Asynchronous Independent Cascades)**: A graph-based model for predicting information diffusion that integrates temporal dynamics, allowing parameters to vary over time. This model outperforms classical approaches in predicting the temporal dynamics of information diffusion. 3. **SONDY (Social Network Dynamics)**: An open-source platform designed to assist researchers and users in understanding social network dynamics by providing functionalities for topic detection, network analysis, and visualization. It aims to facilitate the comparison and evaluation of different techniques for mining social data. The thesis addresses the challenges of identifying popular topics, understanding the paths and mechanisms of information diffusion, and identifying influential spreaders. It also highlights the need for practical tools to support these analyses, which are often lacking in existing research. The work is grounded in the formal representation of online social networks as graphs and the concept of social influence, which can lead to herd behaviors and informational cascades.Adrien Guille's PhD thesis focuses on understanding and analyzing information diffusion in online social networks. The key contributions of this work include: 1. **Survey of Developments**: A comprehensive review of existing methods and models for information diffusion, including popular topic detection and influential spreader identification. 2. **T-BaSIC (Time-Based Asynchronous Independent Cascades)**: A graph-based model for predicting information diffusion that integrates temporal dynamics, allowing parameters to vary over time. This model outperforms classical approaches in predicting the temporal dynamics of information diffusion. 3. **SONDY (Social Network Dynamics)**: An open-source platform designed to assist researchers and users in understanding social network dynamics by providing functionalities for topic detection, network analysis, and visualization. It aims to facilitate the comparison and evaluation of different techniques for mining social data. The thesis addresses the challenges of identifying popular topics, understanding the paths and mechanisms of information diffusion, and identifying influential spreaders. It also highlights the need for practical tools to support these analyses, which are often lacking in existing research. The work is grounded in the formal representation of online social networks as graphs and the concept of social influence, which can lead to herd behaviors and informational cascades.