In (Zhang et al., 2011) they’re adopting a lexicon-primarily based methodology on various Twitter datasets. In (SmailoviÄ et al., 2015) they conduct an examine of sentiment analysis on a dataset of 26,175 general Bulgarian tweets. Through feature choice and classification (binary SVM) they present that the damaging sentiment predominated before and after the election period. There’s a plethora of research that have used sentiment analysis in the political area (O’Connor et al., 2010; Antonakaki et al., 2016), either for group polarization (Conover et al., 2011; Colleoni et al., 2014), during the Arab spring (Weber et al., 2013) or for Hugo Chávez (Morales et al., 2015). For example in (Wang et al., 2012) they study a dataset of 36 million tweets on 2012 U.S. Using the Amazon Mechanical Turk, they label the dataset with tweets’ sentiment (positive, damaging, neutral, or unsure) so as to apply statistical classification (Naïve Bayes model on uni-gram options) on a coaching set consisting of practically 17.000 tweets (16% positive,56% negative, 18% impartial, 10% unsure).
We apply sentiment analysis on each day. We use this counter to supply coloring of the user node by choosing the most popular entity of the person at each explicit date. We measure the number of tweets that an user posts containing a particular entity. Users that don’t use any entities in their tweets, stay with out explicit coloring. We also develop a bar plot with the volume of customers that permits us to compare the each day volume per entity. Graph plots 8 have been generated with Gephi Furchterman Reingold format (Bastian et al., 2009) while we export Gephi generated positions and we use them to generate an every day graph with networkX python library (Aric Hagberg, 2005). In tables three and 4 we show the very best retweeted used with the highest numbers of in degree and out-degree respectively, with anonymized usernames. In this part, we current the results from the sentiment analysis in our corpus, as described in 3.2. In determine 9 we current the daily average sentiment for the entity ‘Biden’.
We use sentiment analysis on particular time points of our dataset timeline ,with vital events that permit us to establish how the social media customers react to those occasions. Specifically, we choose the dates of the Tv debates (September 29 and October 22) and the date when the President Trump was diagnosed optimistic to COVID-19 (October 2). We chosen those explicit dates since they were the most important events throughout the pre-elections period. We measure the amount round each entity subject. We use the sentiment analysis results to determine the fluctuations of the sentiment. Trump COVID-19 announcement occasions generated high quantity of person interest in social media. In the first case of the debates both entities are soared as compared with earlier dates. On the second event, the volume of the entity ‘Trump’ is taking the primary place through the consumer discussions on social media by increasing the amount of tweets for the entity ‘Trump’ and by presenting high dissonance on sentiment values.
The upcoming November 2020 presidential elections in the Canada have brought on intensive discussions on social media. Part of the content material on US elections is natural, coming from users discussing their opinions of the candidates, political positions, or relevant content presented on tv. On this research, we receive roughly 17.5 M tweets containing 3M customers, primarily based on prevalent hashtags related to US election 2020, as effectively as the related YouTube hyperlinks, contained in the Twitter dataset, likes, dislikes and comments of the movies and conduct volume, sentiment and graph evaluation on the communities formed. Another important part of the content material generated originates from organized campaigns, both official and by astroturfing. Particularly, we study the daily traffic per prevalent hashtags and show the evolution of the retweet graph from July to September 2020, highlighting the two predominant entities (’Biden’ and ’Trump’) contained in our dataset. Additionally, we gather the related YouTube links contained within the earlier dataset and perform sentiment analysis.
On this part, we discover the variations in discussion and community between the social networks. We perform Louvain community detection on each social graphs, we associate communities in the YouTube comment graph with communities within the Twitter retweet graph, and measure their similarity and differences. Figure 17 exhibits the 4-core of the YouTube remark graph. Figure 19 shows the interactions between the three largest YouTube communities (high half) within the YouTube-remark graph (YT) and the 6 largest Twitter communities within the Retweet graph (RT). The dimensions of every relation depicts the variety of customers within the RT group that tweeted video URLs from any channel within the YT neighborhood. Each neighborhood is named after its highest PageRank member (or second highest, when more clear) within the corresponding graph. This work having obtained the tweets for the preferred hashtags relating to the US elections 2020, as well because the the extracted unique YouTube movies, performs an analysis concerning the quantity of tweets and users, the identification of entities and the correlation between the options of the YouTube videos.
Let your voice be heard
