Employing a Text Mining Approach for Yellowest Movement A Case Study on Twitter Data

Social media has become a part of people s daily lives. On social networks, users freely publish their opinions and beliefs. Twitter is a social network that publishes millions of messages daily in different languages. Twitter plays an important role in the free flow of information. Analyzing English data gives access to some of the user s opinions and tendencies. To achieve more insight into user, it is necessary to do sentiment analysis for non-English data. In this study, we aim to do data mining on the French and English tweets of the yellow vest. To do this, we created 2 new datasets. We employ machine translation method for sentiment analysis in French tweets. Our results indicate that English tweets have a more negative opinion than French tweets. We found that the most of French tweets are ruled by users that are not very well known also observed that English tweets are ruled by journalists and politicians. We create a lexicon of the English labeled tweets and French translated tweets that are labeled. We perform a sentiment classification for the labeled tweets and achieve an accuracy of 91.32. This study is the first work about yellow vest movement.

Keywords:

Twitter , YellowVest , Natural Language Processing , Machine Translation , Sentiment Analysis , Data science , user influence