Twitter is a leading social networking site when it comes down to topics such as politics, news, and trends around the globe. Another main reason for people to use Twitter is because they are able to share their emotions and feelings with others and form new relationships and views. With about 330 million users on Twitter (in 2020), it continues to rapidly grow, but at the same time, it is also losing users at a fast pace. In 2019, Twitter had 340 million users, but a year later, it lost 10 million of them. The goal of this chapter is to find the reasons of three questions. The first, to find the reason behind Twitter losing its users. The second, to see how a user changes behavior after usage of Twitter, and third, how a user's behavior changes when expanding his/her social circle on Twitter. For all of these questions, this chapter has designed a data set and executed experiments based on the authors' hypotheses. The results report the accuracies of each of these hypotheses.
TopIntroduction
Twitter is a well-known social networking service where users can interact while using Tweets. Users can post images, videos, gifs and even plain text and interact with their followers. Twitter was first established by Jack Dorsey, Noah Glass, Evan Williams, and Biz Stone in March 2006 and later, in July, it was launched. By 2012, there were more than a 100 million users posting 340 million Tweets in a day. Twitter handled about 1.6 billion search queries every day in the same year. It was named one of the top ten most visited websites and in 2013 and in 2018, twitter established more than 321 million active users.
There has been a significant volume of research on analyzing twitter trends (Keikhosrokiani & Asl, 2022). For instance, Branz & Brockmann, (2018), by acquiring data through sentiment analysis, then filtering it, analyzing it, and finally, interpreting it. They collected more than 1,000 Tweets using Twitter API j4 java tool. Their aim is to present an outlook for analyzing sentiment of Tweets towards filtering and interpreting the data. Along these lines, this perspective is affirmed by the two examination inquiries from programming (SE) field. The accompanying theories were proposed:
Theory 1: The quantity of good emotions, in Tweets, about games, was differed essentially between the male and female designers. Guys indicated more good sentiments towards sports related themes when contrasted with females.
Theory 2: The measure of estimation communicated in the Tweets differs altogether between engineers from collectivist societies and architects from independent societies, with clients from the previous communicating less slant.
Sentiment Analysis was performed on Tweets utilizing WEKA. It is utilized to choose the sorts of feelings and their force in Tweets. They applied different feelings like outrage, sicken, dread, bliss, pity, and trust. The assessment scores depend on estimation dictionaries.
The assortment of solo channels was extended to separate highlights for slant order, considering focuses, for example, emoticons, words and articulations demonstrating a specific feeling as well as feeling, and hashtags. The methodology is affirmed by applying it to two exploration questions and this affirmed the previous discoveries on social and sex disparity in opinion articulation.
Grant Williams and Mahmoud (2017) introduced an investigation that was pointed toward identifying, arranging, and deciphering the different feelings of various programming clients' Tweets. They move the thought viewing writing computer programs customers' Tweets instead of originators' Tweets. They show that emotions imparted in writing computer programs structures' with, unravel and legitimize their end-customer's reactions. The data was then normally requested using SentiStrength and two comprehensively valuable substance. The managed content are exact all around valuable presumption assessment procedures in recognizing unequivocal conveyed in programming Tweets. The work portrayed affirmation and thought examination assessment feeling in programming Tweets.
Based on other reports and research, the reasons for people leaving Twitter is because people were annoyed with the application, or they forgot about it after making an account, or also that they could get the same information that they might have wanted through other apps, etc. There are other various reasons in different papers, and we will go through them to come to our own conclusion.
The motivation for this work is to analyze users’ behaviors on Twitter. The research questions that would be answered here are:
- i)
What leads the users to leave Twitter? (Problem 1)
- ii)
How a user’s behavior changes while using Twitter? (Problem 2)
- iii)
How a user’s behavior changes while expanding his/her social circle on Twitter? (Problem 3)
Rest of the sections of this chapter presents the proposed work. The next section presents the literature review. This is followed by methodology and hypothesis formulation. Then the results are presented. The paper concludes with discussion and future work.