Subject: Statistics

This project will extend the analysis conducted by Seri, Rocci, and Murphy to investigate the performances of rugby players in major leagues using a soft clustering approach (https://link.springer.com/article/10.1007/s00180-025-01655-w). Rather than assigning each player to a single performance category, soft clustering will allow players to be associated with multiple performance profiles to varying degrees, based on a range of performance indicators. This approach aims to provide a more nuanced characterization of players and to identify similarities and differences in their performance patterns. Data are anticipated to be sourced from available online repositories, e.g https://github.com/seanyboi/rugbypy. Required skills: Statistical Inference, Statistical Modelling

Building on the approach introduced by Basini, Tsouli, Ntzoufras, and Friel (https://academic.oup.com/jrsssa/article/186/3/530/7035692#409156010), this project aims to assess competitive balance in major men’s volleyball leagues and competitions (e.g., the Volleyball Nations League) across multiple seasons. The analysis will investigate patterns in team competitiveness and how these evolve over time. Data will be sourced primarily from the Volleyball World website. The statistical analysis will employ the extension of the Stochastic Block Model (SBM) developed in the aforementioned paper. Required skills: Statistical Inference, Statistical Modelling

This thesis investigates cause-specific mortality trends in Finland using data from Statistics Finland (https://stat.fi/en). Age- and sex-specific mortality rates for a major cause of death, such as cardiovascular disease, will be analysed using the Lee–Carter model, which separates mortality patterns into age and time components. The model will be used to examine how mortality has changed across age groups and between men and women, and to produce forecasts of future mortality. The accuracy of these forecasts will be evaluated using out-of-sample data. Required skills: Statistical Inference, Statistical Modelling  

This thesis investigates patterns in members of the European Parliament (MEP) voting behaviour using data from the European Parliament Vote Monitor (EPVM) (https://epvm.iep.unibocconi.eu/). Clustering methods such as k-means, hierarchical clustering, and model-based clustering will be used to identify groups of MEPs with similar voting patterns. The resulting clusters will be compared with characteristics such as political group and country to assess whether voting behaviour follows existing political divisions or reveals alternative groupings.  Required skills: Statistical Inference, Statistical Modelling 

The organization VoteWatch collected data for the European Parliament (EP) over the years 2014 – 2019. The data contain multiple information on the items being voted by the members of parliament (MPs) over that period, such as the area of a policy, the legal classification of the document being voted, etc. The aim of this … Read more