摘要：In this article, we perform sentiment analyses of Twitter location data. We use two case studies: US presidential elections of 2016 and UK general elections of 2017.For US elections,we plot state-wise user sentiment towards Hillary Clinton and Donald Trump.For UKelections, we download two disparate datasets, using keywords and location coordinates,looking for similar tendencies in sentiment towards political candidates and parties.Theoverall objective of the two case studies is to evaluate similarity between sentiment of location-based tweets and on-ground public opinion reflected in election results. We discoverTwitter location sentiment does indeed corroborate with the election result in both cases.Wealso discover similar tendencies in Twitter sentiment towards political candidates and partiesregardless of the methodology adopted for data collection.
