Estimating sentiment in social media : a case study of the migrant caravans news on Twitter

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Shravya Muttineni (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site:
Jing Deng

Abstract: Social media has empowered us to gain immense knowledge, criticism, and positivity from every corner of the world. Accessing social network sites like Twitter and Facebook has slowly become the norm, as people are now inclined towards social media for news and opinions rather than reading traditional newspapers or watching TV news. There are many instances when a piece of news in these sites, irrespective of its credibility and reliability, has a significant impact on people’s opinions towards it. Social Media has created a platform for many individual users, small business owners and large co-operations to make a living by targeting their users and generating leads through digital Marketing strategies. Although these networks being extremely useful in many ways from the recent events we have seen a greater influence of these platforms in one’s life effecting mental well being and creating a social dilemma. In this work, we focus on identifying such problems of social media posting sentiment efficiently with the use of estimation and references through the methods of sentiment analysis for tweets. We calculated the sentiments of all the tweets related to the migrant caravan issue. Using the scores from these tweets, I have analyzed the outcomes. We used Vader Sentiment Analysis to analyze the sentiment of each tweet before estimating the sentiment on Twitter toward this news. We used different weights for the tweets in particular weeks. Estimated the effects it could have on the users from the computational analysis we performed. we came through different kinds of visualizations to present out work. we even tried some machine learning algorithms targeting better efficiency and automated results but, left most of these for our future work. We investigated some queried tweets of around 40 weeks of a sensational tragic issue of “Central American migrant caravans” and how the tweets played their critical roles in the overall sentiment on the twittersphere. We show that such simple estimates can be very accurate. Our study can be helpful in its efficient estimation of social media content sentiment analysis. This includes potential identification of user cluster and inherent communities, dynamic community detection on-the-fly, community structure migrations, etc.

Additional Information

Language: English
Date: 2021
Social media, Twitter, Migrant caravans, Vader Sentiment Analysis, Central American
Sentiment analysis
Mass media and public opinion
Immigrants $x Public opinion

Email this document to