Research Area:  Data Mining

Abstract:

Analyzing public information from social networking sites could produce exciting results and insights on the public opinion of almost any product, service, or behaviour. One of the most effective and accurate public sentiment indicators is through social networks data mining, as many users tend to express their opinions online. The internet-s advanced technology has managed to increase activity in blogging, tagging, posting, and online social networking. As a result, people are starting to grow interested in mining these vast data resources to analyze opinions. Sentiment analysis is one of the computational techniques of opinion, sentiments, and the variety of texts subjectivity. In this paper, the methodology of determining these public opinions are discussed. The development of a program for sentiment analysis is done to create a platform for social network analysis. This paper also discusses the sentiment analysis design, gathering data, training the data, and visualizing the data using the Python library. Finally, a platform is designed in order for other users to search the sentiment results of particular topics of interest. A total of 3000 Reddit data and 3000 Twitter data has been gathered, cleaned, analyzed, and visualized in this research. The analysis has produced an excellent percentage result of 83% and 77% for Twitter and Reddit data, respectively. Moreover, the GUI platform has been built using the Tkinter library.

Keywords:  
Social network analysis
sentiment analysis
Twitter
Reddit
Python

Author(s) Name:  Teddy Surya Gunawan; Nur Aleah Jehan Abdullah; Mira Kartiwi; Eko Ihsanto

Publisher name:  IEEE

DOI:  10.1109/CITSM50537.2020.9268866

Volume Information:  

Paper Link:   https://ieeexplore.ieee.org/abstract/document/9268866