Projects in the area of Analysis of Online Social Networks

These research projects aim to explore (1) new methods for quantitative measurement of user opinion and (2) models for prediction of user behaviours on online social networks, such as X (formerly known as Twitter). 

A wide range of scientific methods can be utilised, including machine learning methods (such as decision tree models, feature importance analysis, neural network models etc.), statistical/computing methods (such as Bayesian inference, dimension deduction algorithms and numerical simulations), and network science methods (such as network community detection and PageRank).  

Two Research Topics are available:

1). New methods for quantitative measurement of user opinion. How to measure a user' opinion based solely on its social connections with other users? We are exploring new methods, for example neural network embedding, to infer and quantify individual's opinion based on their social networking data only. Please read the previous BSc/MSc reports below to get a better understanding of the project. 

• BSc project thesis (2024) Title: Opinion Measurement Based on Social Connectivity Data 

• MSc project thesis (2024) Title: Ideological positioning of US Congress members using Twitter following networks

2). Models for prediction of user behaviours on online social networks. How to predict the probability of a user reposting a message posted by another user? While most existing methods focused on analysing the content of message, our experimental results suggest that a significant component of reposting behaviour can be predicted based on users' profile and past behaviour, and is independent of the content of messages. 

Requirements: 

• Good knowledge and skills in mathematics and machine learning. Knowledge on network science is a bonus. 

• Good skills and experience in programming, such as Python.

Contact:

Please contact Prof. Shi ZHOU <s.zhou (at) ucl.ac.uk> for more information.