SCIS 2060 - Social Network Analysis for Social Media and Politics

A number of research works across several disciplines rely on social media data. Because of its availability and granularity, studies in political sciences, sociology, journalism, and economics leverage social media data for empirical analysis and experimentation. Beyond social media studies, the methodological approaches used in these studies have been popularized and applied to survey-based research, analysis of political discourse, political competition, and news media, to name a few. In this course, we will overview social media data analysis, structured along three main thematic axes that have contributed to the understanding of online informational, social, and political ecosystems: segregation (linked to isolation and so-called echo chamber phenomena), diversity (linked to news media diets and evaluation of algorithmic recommendations), and polarization (linked to political competition, extremism, and opinion dynamics). Through the analysis of these three types of phenomena, this course provides students with 1) a methodologically structured overview of the landscape of social media studies in scientific literature, and 2) a conceptually robust theoretical understanding of the underlying models that are leveraged in these studies, and that can be applied to a wide range qualitative and quantitative studies in computational social sciences in general. This course is organized around two types of activity: 1) presentation and collective analysis of different studies leveraging a wide but compactly organized set of models, tools, and theories, and 2) hands-on practical coding experiences and techniques using real-world social media data through the development of group projects. Previous experience with data analysis in Python is required, but the course is conceived so that students with different coding skills may participate and improve their skills. The main goal of the course is for students to boost their ability to leverage network data in research projects. Requirements: students taking this course should be familiar with loading and treatment of CSV data files on Python using pandas. Basic experience using statistical analysis modules such as scikitlearn is desirable but not required. No previous knowledge of network data is required.

Pedro RAMACIOTTI MORALES

Séminaire

English

Day 1 (Monday) Morning (9:00 – 12:00) Afternoon (13:00 – 16:00) - Introduction: topics, scope, organization, and evaluation - Network models for online interactional data traces - Heterogeneous web entities and interactions - Similarity networks - Hands-on network data manipulation and datasets - Political network analysis overview: network diversity, segregation and polarization - First groupwork session and presentation of project: creation of groups, exploring datasets for the project Day 2 (Tuesday) Morning (9:00 – 12:00) Afternoon (13:00 – 16:00) - Network diversity: news media diets, diversity of recommendations, social diversity in networks - Analyzing algorithmic recommendations in social media - Guided groupwork session: - Early project idea presentations Day 3 (Wednesday) Morning (9:00 – 12:00) Afternoon (13:00 – 16:00) - Conceptualizing and measuring segregation in networks - Measuring segregation - Community detection in social networks - Conceptualizing and measuring polarization in networks - Ideal point estimation and spatialization of social media data Day 4 (Thursday) Morning (9:00 – 12:00) Afternoon (13:00 – 16:00) - Collective analysis of a Twitter case study involving segregation, diversity, and polarization - Guided groupwork session - Sprint session (part 1) Day 5 (Friday) Morning (9:00 – 12:00) Afternoon (13:00 – 16:00) Sprint session (part 2) - Finishing slides for group presentations - Group presentations

Intermediate experience coding in Python (see details in description)

Spring 2023-2024

Attendance is mandatory Evaluation: The evaluation of this course consists of a group project to be developed through the week and presented Friday afternoon. In this short project, students will select datasets available online to propose a political analysis involving either algorithmic recommendation, news media consumption, social networks, online cultural consumption, or some combination of these and other related online phenomena.