This paper analyses the 2016 EU Referendum results, correlates the results with the 2015 General Election results and aims to further uncover discrepancies in London. In London, the Referendum was organised into 33 voting regions, whereas there are 73 Westminster constituencies for the General Election and 637 wards which provide additional data of interest. This creates a problem when attempting to find insight without losing information and attempts have been made to address this issue and also how the results are subsequently impacted.
The Referendum and General Election results are correlated and compared. Using the relationship models which have been created in an attempt to discover how individual constituencies and local wards in London voted in the Referendum. Accuracy and performance of the model is reviewed and suggestions on how to improve results are made for future work.
Python, R and Tableau have been used to organise, visualise and apply regression analysis to generate estimates for the referendum voting patterns in smaller regions of London.