Brexit and Uncertainty: an empirical and dynamic analysis of an event through Taxonomies and Twitter Data.

In this Thesis we analyse data coming from approximately 115,000 UK Tweets from 05/2018 to 05/2019, whose keyword is the term "Uncertainty". Our final aim is to deepen our understanding of how uncertainty is perceived in different geographical areas, and how people link "Brexit" to Uncertainty variables. In order to do so, we use innovative methodologies such as taxonomies, which we intentionally designed around economic variables of interest for our analysis. These are released as open source for further research purposes. Our work is divided in three parts. In the first one, we conduct a comparison between aggregated and disaggregated analysis by geographic area, to show differences in perception of uncertainty in different parts of the Kingdom. The second part focuses on the sentiment analysis by geographic area, using two different measurement techniques: classical sentiment analysis and NRC Emotions Lexicon. Lastly, we devote our attention to co-occurrence matrices, that we enrich with 3 levels of network analysis.