Abstract: This project is about analysing social media data around the elections with a particular focus on the polarisingtopic of xenophobia. The goal is to gain a better understanding of the political discourse that may fuel xenophobia.Twitter data sets from municipal and national elections from 2016 onwards are available. By analysing Twitterdata, we can quantify polarisation regarding xenophobia, track any changes over time and possibly map thetweets and conversations to understand their impact on the elections.Refer to the document ”Group 15 Rockers Modelling MIT808.pdf” for information about the modelling choicesthat were implemented and that lead to choosing the Distilled RoBERTa-base model .